Deeplinks

The internet is a vast and confusing place no matter your age. Trying to help youth engage with it safely and privately is a complicated topic that touches on many different skillsets and ideas. But just because it’s complicated doesn’t mean it’s impossible. We’ve launched our Online Privacy for Parents and Youth Surveillance Self-Defense guide in hopes to provide a framework for finding solutions that work for you. 

EFF rolls out this crucial resource as state and local governments consider ill-advised laws that cut young people off from essential information and experiences. This particularly harms vulnerable youth and marginalized groups who often find safety in supportive online communities they can't access offline. Such laws also entail age gating, which compromises the privacy of all internet users and stifles young users’ ability to learn, grow, and express themselves.

Young people have unique needs and those needs change over time. Trying to protect from everything all the time is an exhausting process, but the security community has long had a solution for this: security planning (also called threat modeling). The idea is to provide a series of questions designed to help you focus on what’s truly important. When it comes to online privacy, this is as applicable for youth and parents as it is for anyone else. 

We used a security plan as our foundation for this guide because deciding what is and isn’t important is the first step we should all take when assessing our privacy risks. But this isn’t a place for quick fixes, and every young person has their unique needs and considerations. There is no single tool or technology fix that can substitute for spending time learning, and then teaching a young person, how the internet works, how to use it, what sorts of harms exist, and how to navigate those harms.

Solutions are rarely about the tools or software you use. It’s about understanding the threats and concerns you have, and then determining what steps you can take to mitigate those risks. We hope our Online Privacy for Parents and Youth guide can help provide the outline to get the conversation started.

Donate Today!

An open internet where people can freely build, share, and remix each other's work is a powerful engine of grassroots innovation we’ve all benefited from. In recent decades, however, corporations have taken control of an increasing share of this work, giving them a greater ability to stifle challengers and limit our choices. When regulators address this by mandating that  tech giants play fair by requiring interoperability, the gatekeepers respond with a tired refrain: Playing fair puts users’ security at risk.

Unfortunately for these trillion-dollar companies, this gotcha isn’t the end of the conversation. Regulators can ensure that meaningful choice and security thrive.

Lawmakers have been catching up, recognizing that the best way to rein in big tech is to require these oligopolies to open up their platforms to independent developers. As we explained six years ago, mandating interoperability helps prevent the misuse of people’s data and enables more privacy-respecting alternatives. Tech giants, in contrast, have been arguing that our data is safer when they have exclusive control.

The European Union’s Digital Markets Act (DMA) put interoperability into law under Article 6(7): “gatekeeper shall allow providers of services and providers of hardware, free of charge, effective interoperability with” a designated Core Platform Service they operate. The language describes an unambiguous requirement to remediate vendor lockouts and anti-competitive behavior of the major service operators—it is not optional.

Apple claimed that enforcing this obligation in mobile app stores would put their customers at risk and do irreparable harm to trust in their brand. It’s an argument reminiscent of Apple’s fight against the right to repair, and in both instances the company is fighting to maintain its restrictive ecosystem to stop people from finding better deals elsewhere. Apple has long cultivated a more corporate-controlled experience with the iPhone, but this argument is now also being used for the relatively-more-open Android devices from Google.

In the same year Apple argued against the DMA, Epic Games sued Google for its restrictive terms of service and extractive policies for revenue generated in apps. Google predictably alleged that opening up its platform to other app stores it couldn’t vet would allow harmful apps to be installed. The case was ultimately decided this year in the U.S. federal appeals court for the Ninth Circuit, with the court rejecting this argument, dealing a blow to  Google’s feudal security arrangement, and paving the way for third-party app stores to be delivered as apps via the Google Play Store.

Policymakers cannot take tech giants’ claims at face value. Companies have every incentive to resist interoperability requirements, because interoperability lowers practical barriers and reduces switching costs for consumers adopting competing services. 

Even where companies offer credible claims from internal technical and security experts, independent experts rarely have the access and information they need to test those claims. Companies have selectively made arguments that security is harmed by interoperability, often in contexts which are to their advantage. These companies are happy to let a general misconception flourish which frames security as inherently opposed to interoperability.

This narrative is wrong. We’ve seen that when interoperability and information security are both set as essential and non-negotiable criteria, secure interoperable systems result—systems that don’t sacrifice the rich features we’ve come to expect of the services we rely on. 

The cross-platform rollout of Rich Communication Services (RCS), and now Message Layer Security (MLS) on top of it is a perfect example of this, creating end-to-end secure messaging by default between Android and iOS. For RCS and MLS, Google was conveniently on the other side of the interoperability argument, campaigning Apple to adopt the standard to overcome the diminished “green bubble” experience iPhone users had when texting non-iPhone users. Once Apple finally agreed to interoperate by implementing common standards, all users benefited from improved messaging security. In cases where legitimate security concerns are raised, these can be assessed and remedied on a case-by-case basis, without sacrificing user agency.

Retrofitting interoperability into already secure closed systems shouldn’t be done carelessly: it requires regulators strike the right balance of pressure and leniency to allow the time and precision necessary to make these renovations. Any components developed with the assumption of privileged access by trusted parties (e.g. your browser engine) are particularly vulnerable to serious security risks when opened up to unaccountable third parties.

By adapting interoperability and information security as essential design principles, we can expect more safety, but also more freedom.

With regard to the DMA, when security concerns are legitimate, we urge the European Commission to work with companies and outside experts to offer guidance in technical implementation and provide clarity on regulatory compliance. This approach paves a path forward that balances the interest of legal oversight and the needs of the companies alike. 

Policymakers wishing to uproot these unfair platforms of control must do so with care and limit the impact this disruption can have. But security concerns must not act as a lever to indefinitely delay compliance. Closed systems should be impartially evaluated, and if concerns are valid, regulators can allow more time and guidance to be given to companies implementing these changes. One possibility is a tiered approach: offering leniency during an initial period where systems need to be retrofitted to abide by new requirements, followed by a subsequent and ongoing stricter observance of fair play.

Once interoperability and information security are adopted as essential design principles, further development with these rights-enhancing principles in mind will more easily follow. The result will be an internet the world deserves—where we can expect more safety, but also more freedom.

The Brazilian National Data Protection Agency (ANPD) plays a pivotal role in the regulation of digital platforms according to new legislation approved in Brazil. Throughout this year, EFF has presented comments, filed submissions, and met with the agency to emphasize crucial guardrails that the enforcement of new regulations should ensure. 

ANPD is responsible for detailing and applying obligations set in Law n. 15.211/2025, which established rules for the protection of children and adolescents in the digital space. While some provisions establish important privacy safeguards, like prohibiting advertising based on profiling of underage users, the law created mandatory age assurance mechanisms for accessing digital products and services. 

We have raised alarms on the privacy and free expression issues around age-gating the internet. Age verification technology is often inaccurate and privacy-invasive. Examples also show how it’s prone to enforcement overreach, resulting in sites blocked beyond legal requirements. In the UK, providers disproportionately restricted several counseling, support and mental health sites under the Online Safety Act. Current age assurance methods risk wrongly walling off adults with no valid IDs, facial differences, physical disabilities, among others. 

We highlighted these concerns in our comments and submissions underscoring that ANPD should avoid a blanket approach to age assurance obligations. Following the legislation, restrictions should apply only to products and services that are inadequate or prohibited to children and adolescents according to categories already set in regulation, with robust protections against subjective and overbroad interpretations. Guarantees such as access to information, freedom to participate in society and meaningful access to technologies also compound what constitutes the best interests of the child.  

Moreover, we stressed that zero knowledge proofs alone fail to ensure strong user privacy, including safeguards that the law itself sought to secure. Our submission explains why other requirements, such as unlinkability, selective disclosure, batch issuance and processing on device, are crucial for achieving greater levels of protection and mitigating privacy harms. Minimum information being shared to the verifier and no “phoning home” to the issuer of a user’s accepted credentials are conditions that prevent actions like collusion and data sharing about the user. 

We also shared inputs about the scope of the law, asking more clarity on which providers are covered by its obligations and the importance of treating differently non-dominant players like federated networks and free software applications. Without that clarity, this could negatively impact beyond gaming and major platforms that are the main target of these regulations. We elaborated on the agency’s regulation of prevention duties, the relevance of modulating them according to the type and structure of providers, and alerted them that their interpretation should not result in general monitoring obligations. 

On another but related front, we sent comments regarding Brazil’s new internet intermediary rules. Here too, ANPD has a critical regulatory role after a Supreme Court’s decision and presidential decrees changed the liability regime initially set in the law known as “Marco Civil.”  

With the turn to notice and takedown mechanisms and duty of care obligations, we emphasized measures to prevent censorship of protected content. These involve ensuring solid means for appeals of restricted posts, avoiding overreliance on AI-based content moderation, strictly tailoring the application of concepts like duty of care and systemic failure, and not normalizing the blocking of platforms and websites.  

We raised concerns about a trend reflected in the new regulation requiring platforms to proactively report user to authorities with unclear guarantees. Finally, we urged the agency to advance essential transparency and accountability measures for enabling greater public oversight over the application of the new regime. 

As we said before, platform accountability under the new rules ultimately hinges on how accountable its enforcement will be by platforms and state institutions. Rigorous commitment to fundamental rights, including freedom of expression, privacy and data protection, is required so that intended protections don’t instead lead to harms. ANPD must uphold this commitment and make sure that new regulations don’t backfire. 

Police agencies across the U.S. have deployed Automated License Plate Readers (ALPRs) to capture images of passing vehicles along with time, date, and GPS coordinates. While the technology itself isn’t new (ALPRs have existed in various forms for more than 50 years, and in the U.S. since 1998), the surveillance landscape has shifted dramatically. Rapid advances in artificial intelligence and national data sharing have transformed police ALPRs into an interconnected, nationwide tracking network. It is pitched to city councils by law enforcement and private vendors like Flock Safety, Motorola Solutions (Vigilant), and Axon as an essential tool for solving violent crime.

In reality, it is an inherently dangerous infrastructure of mass surveillance.

Since 2012, EFF has worked with others to expose the severe digital rights harms of ALPR systems. For example, we’ve recently documented how these camera networks allow law enforcement to search for immigrants, track peaceful protesters, target Romani communities, and track women seeking reproductive healthcare. Police also routinely use ALPRs for low-level matters like enforcing school residency requirements or investigating noise complaints. Many officers also use ALPR databases to stalk past romantic partners. Recent EFF analysis revealed officers using the system without legitimate justification—logging search reasons like "LOL" and "LMAO."

Understandably, the public backlash to police ALPR technology is accelerating, and state lawmakers across the country are taking action.

New Hampshire enacted the nation’s first state law on ALPR use in 2007, but technological advances and misconduct have driven even more recent legislative action. In 2026 alone, six states enacted laws that limit police use of ALPRs: Connecticut, Kentucky, New Jersey, New Mexico, Oregon and Washington.

This is a good start, but far more is needed. This blog presents a framework for state legislators who want to stop ALPR mass police surveillance:

1. Just ban it. It is incompatible with a free society and cannot be fixed.
2. In the absence of a flat ban, reduce its harms. Most importantly: limit police use of ALPRs to real-time comparison of passing plates to hotlists; require a warrant to place a plate on a hotlist; impose strict limits on the creation and use of those hotlists; and prohibit the storage of ALPR data (and thus prevent the subsequent search and disclosure of that data).
3. In the absence of a prohibition on storing ALPR data, reduce its harms. Most importantly, place strict limits on police storage, search, and disclosure of ALPR data.
4. Ban three especially dangerous police practices: placement of ALPRs near sensitive locations like hospitals; purchase of ALPR data; and use of ALPR data to try to predict crime.
5. Require transparency about ALPR mass surveillance, including: a 30-day public notice before a city procures police ALPRs; a statewide registry of all police ALPRs; and regular, public audits and statistical reports about police usage.
6. Ensure effective enforcement of all new ALPR restraints, including by means of a private right of action and the exclusionary rule.

These policy battles often unfold locally, so state legislators must ensure that state law is a regulatory floor, never a ceiling. State preemption should never force unwanted camera networks onto communities that have chosen to reject them. Local government must retain authority to enact full bans, deny funding, reject contracts, and decommission existing camera networks.

• The Gold Standard: Statewide Bans on ALPR Mass Surveillance
• Strict Limits on Real-time Use of ALPRs 
  • Limits on creating hotlists
  • Limits on using hotlists
• Strict Limits on Storage, Search, and Disclosure of ALPR Data
  • Storage limits 
  • Search limits 
  • Disclosure limits 
• Transparency 
• Additional Restrictions
• Enforcement 
• Conclusion

As in our previous post on ALPR policy, our focus here is ALPR mass surveillance: the continuous dragnet collection and retention of location data on innocent drivers. Different issues are raised by automated traffic enforcement (ATE) systems, such as standalone red-light or speed cameras that solely photograph and ticket specific traffic violations, without building a searchable network about uninvolved drivers. As such, state laws against ALPR mass surveillance should not cover ATE. This is the approach in at least six states: Connecticut, Minnesota, North Carolina, Oregon, Virginia, and Washington. The most common approach is to exclude ATE from the law’s definition of “ALPR.”

Of course, ATE carries risks and demands legal restrictions too. This includes tight limits on data collection, retention, usage, and disclosure—all of which should be addressed in a different statute. Lawmakers must also actively guard against vendor efforts to piggyback mass surveillance tools onto ATE contracts.

Statutory limits cannot make ALPR mass surveillance safe. Because the technology relies on the indiscriminate recording of every driver’s movements, the risks are baked into the system itself. Also, innocent motorists will be menaced by police due to false matches, bad actors will misuse location data, and sensitive personal movements will be exposed to broad government overreach. Promising harsher penalties for officer misuse cannot fix an inherently flawed architecture. Thus, state lawmakers should pursue a complete, statewide ban on ALPR mass surveillance.

First, lawmakers should prohibit all state-level law enforcement agencies from acquiring, retaining, accessing, or using ALPRs or any data derived from them.

Second, they must impose the same prohibition on county, municipal, and other local law enforcement agencies.

These prohibitions should require any agency ending an ALPR program (both state and local) to fully decommission the programs within a strict timeframe. This includes requiring agencies to: physically remove all camera hardware; revoke all user, network, and database access credentials; permanently delete all stored location data; and require vendor certification of compliance.

Third, ban the use of state funds for state or local agencies to acquire, maintain, or access ALPR systems. This must include direct appropriations, law enforcement grants, pass-through funds, and state-administered federal funds. In August 2026, Texas Governor Greg Abbott banned state agencies from spending public funds on Flock cameras.

Fourth, ban state and local agencies from installing or operating ALPRs on state-owned highways, roadways, or public properties. In August 2026, Florida’s Department of Transportation  ordered the removal of all ALPRs from the right-of-way on state highways within 30 days. The order revokes all previously approved permits to install ALPRs, and bars future permits.

An outright prohibition is the only effective way to eliminate this mass surveillance. This follows the lead of municipalities across the country that have already enacted bans. However, where there’s political resistance, state lawmakers might consider a moratorium that prevents further expansion of ALPR networks.

In the absence of a ban on police ALPR mass surveillance, state legislators must restrict ALPRs to real-time comparison of passing plates to hotlists. ALPR data is simply too dangerous for police to safely store, search, and disclose. An effective way to do this, as in New Hampshire, is to require deletion of ALPR data within three minutes after collection. That means no ALPR data is stored after comparison to a hotlist, and thus that no ALPR data about unmatched cars can be searched or disclosed.

A state legislature must also reduce the harms of police use of real-time ALPRs by strictly limiting the creation and use of law enforcement hotlists.

Roughly speaking, there are two kinds of police hotlists. An agency might link its ALPRs to external hotlists created by another agency, such as the FBI’s National Crime Information Center (NCIC) or a state’s department of motor vehicles. Also, a police agency might link its ALPRs to its own internal hotlist, created by its own officers, of people under investigation. State legislators must strictly limit both.

First, limit the kinds of external hotlists that police agencies may link to. For example, ban police agencies from linking to the NCIC’s “immigration violator” hotlist, which ICE uses to hunt immigrants with the ALPRs operated by police agencies across the country.

Second, if a police agency links to external hotlists, require the agency to update that link every day. Otherwise, drivers may be detained even though they have been removed from the list. This is required by Montana, Nebraska, North Carolina, and Virginia.

Third, forbid officers from adding a plate on a hotlist, absent a warrant from a judge based on probable cause and particularity. Without this judicial check on police power, an officer stalked a potential romantic partner by placing her plate on a hotlist.

As with other laws requiring a warrant (such as the federal Wiretap Act), there should be an exception where: (1) there is an emergency involving immediate danger of death or serious bodily injury to a person; (2) there is no time to acquire a warrant; and (3) there are grounds upon which a warrant could be granted. This exception might apply, for example, when police have probable cause to believe a child has just been abducted. In such cases, the statute must require police to seek a warrant within 48 hours, and if it is denied, to immediately remove the plate from the hotlist.

Minnesota requires a warrant or an emergency to use ALPRs “to monitor or track an individual who is the subject of an active criminal investigation.” In Maine, an ALPR system may use only information based on reasonable suspicion.

Fourth, forbid officers from adding a plate to a hotlist, except in connection with a serious violent crime. This would include investigation of a murder but not pursuit of a traffic scofflaw.

Fifth, forbid officers from adding a plate on a hotlist based, in whole or in part, on: a person’s race, ethnicity, national origin, religion, immigration status, disability, sex, gender identity, or sexual orientation; their political beliefs, associations, or participation in First Amendment-protected activity; or their pursuit of reproductive, gender-affirming, or other health care. Connecticut and Washington forbid use of ALPR systems on these grounds, and Virginia forbids tracking a person based on their protected speech.

Sixth, require officers to log their entry of a plate into a hotlist. This must include the officer’s name and agency, the date of the entry, a narrative description of the reasons for the entry, any case numbers, and an upload of the warrant authorizing the entry. Oregon requires an officer to log “the specific law enforcement purpose of the entry,” and the officer’s name.

First, if the plate of a passing car does not match a hotlist, laws must require immediate deletion of all ALPR data about that car. New Hampshire state law requires deletion within three minutes. This is the most effective way to limit ALPRs to real-time identification of vehicles and drivers of interest to police, as opposed to amassing a dangerous database of the comings and goings of the entire driving public, available later for retrospective search by police.

Second, if the plate of a passing car does match a hotlist, then limit storage, search, and disclosure of its ALPR data to what is strictly necessary to the investigation that placed the car on the hotlist.

Third, require police officers, before stopping a car because of an apparent real-time match, to visually and independently confirm that the car’s plate number, plate issuing state, and other vehicle characteristics in fact do match the car on the hotlist. Many innocent motorists have been detained and menaced by police because of ALPR false positives. Confirmation is required by Kentucky, Montana, New Hampshire, Oregon, Virginia, and Washington.

Again, state legislatures should simply ban police use of ALPRs, or at least police storage of ALPR data. Third best is harm reduction through strict limits on storage, search, and disclosure of ALPR data.

First, limit the storage period for ALPR data to days, not months, and then require deletion. More than 99% of the drivers who pass an ALPR are not on a hotlist. Tracking their movements anyway, just in case they are later relevant to a police investigation, is criminalization of the entire driving population. Thus, four states already require deletion after 21 days: Connecticut, Maine, Virginia, and Washington. Even Flock has reduced its default retention period to seven days.

By “delete,” we mean permanently erase it, so it is impossible for anybody to access it again. It is not enough to merely store it in a way that makes it harder to access, such as archiving it, or transferring it from one database to another. In the words of the Virginia statute: “System data shall be purged … in such a manner that such data is destroyed and not recoverable.”

Second, limit the types of ALPR data that can be stored: just a photo of the plate, the AI-based read of the plate number, the date and time, and the location. Conversely, ban retention of photos of people in the car, and of features of the car beyond the plate as well as AI-based reads of those photos (respectively, “facial recognition” and “vehicle fingerprints”).

Many states have enacted such rules. New Hampshire bans ALPRs from “photographing or recording or producing images of the occupants of a motor vehicle.” Missouri bans integration of facial recognition into ALPR systems. New Hampshire and Montana limit ALPRs to “the sole purpose of recording and checking license plates.” Minnesota provides that “data collected by an [ALPR] must be limited to” plate numbers; date, time, and location; and pictures of plate, vehicle, and surrounding area.

First, laws should ban “geofence” searches, which are searches in which police use ALPRs to identify all of the drivers present at a particular time and place, based on police claims that one of them might be the suspect they are looking for. Even a warrant cannot justify such general searches, because the Fourth Amendment requires warrants to “particularly” describe the person or car being investigated. These dragnet searches of stored ALPR data are the modern-day equivalent of the colonial-era “writs of assistance,” against which our founders fought a revolution to end.

Second, laws should forbid officers from searching stored ALPR data, unless they first acquire a warrant from a judge based on probable cause and particularity. The constant revelations of bad police searches of ALPR data demonstrate the need for—at the very least—judicial control over this dangerous surveillance technology. A federal judge in Oklahoma recently held that the constitution requires this judicial control, because ALPR mass surveillance invades reasonable expectations of privacy. EFF has filed lawsuits making this demand in California and Wisconsin, along with local ACLU affiliates: SIREN and CAIR-CA v. San Jose, and Wisconsin Council of Churches v. Milwaukee, Racine, Kenosha, and Green Bay.

Three states (Minnesota, Vermont, and Virginia) require reasonable suspicion of a crime before searching ALPR data, and a fourth state (California) requires this from its highway patrol. This is a step in the right direction, but not enough: the decider must be a judge, and the test must be probable cause.

Third, officers should be prohibited from searching stored ALPR data, unless they are investigating a serious violent crime. On the flip side, ban searches for immigrants, protesters, or traffic scofflaws.

Some states have moved in this direction. Missouri limits use of ALPR data “criminal justice” and “lifesaving efforts.” North Carolina bans its use for traffic enforcement. Connecticut bans its use based on race and other protected classes, to identify First Amendment activity, for immigration enforcement, or because of access to reproductive or gender-affirming health care. Idaho bans its use for political activity.

Fourth, police officers must be required to log their searches of stored ALPR data concerning a particular car. This must include the officer’s name and agency, the date and time of search, a narrative description of the reasons for search, any case numbers, and an upload of the warrant authorizing the search. The officer must write the narrative in their own words. A drop-down menu of reasons is no substitute, because it is too easy for officers to dodge an explanation of the actual reasons by clicking a pre-written “reason.”

Many states require officers to log their searches of ALPR data, including California, Connecticut, Idaho, Maryland, Missouri, Montana, New Hampshire, Oregon, and Virginia. In Virginia, this must include an officer’s documentation of both their “specific purpose” to search, and the “associated call for service or case number.”

The “disclosure” limits below must broadly address sharing from an agency that collected ALPR data to another entity. This includes an agency opening its database to access and search by another entity. It also includes an agency transferring data from its database to another entity’s database.

First, ban disclosure of ALPR data to any federal or out-of-state agency. Federal and out-of-state agencies often lack the necessary protections of the driving public. Connecticut forbids disclosures to the federal government or agencies in 46 states, unless they have a warrant. California forbids disclosures, unless otherwise permitted by law.

Second, ban disclosure of ALPR data to another agency, unless it follows all of the rules above for an agency’s search of its own data (that is, the ban on geofence searches, the requirement of a warrant, the limitation to investigation of serious violent crimes, and the requirement of an audit trail).

Some states have taken steps in this direction. Utah requires, as a condition of disclosure to a government entity or a defendant in a criminal case, a court order finding reasonable suspicion of crime. Washington forbids disclosure “except as required in a judicial proceeding.” Minnesota and Virginia require the disclosing agency to find reasonable suspicion. Minnesota and Connecticut forbid disclosure unless the receiving agency follows the rules applicable to the disclosing agency.

Third, place strict limits on participation in a multi-agency disclosure network, such as Flock or Vigilant. Connecticut forbids participation “in a system or network that shares” ALPR data, unless it requires all participants to affirm in writing that they will not use or re-disclose that state’s data except in compliance with that state’s laws. Connecticut also forbids “bulk” data disclosure, and Maryland forbids data “upload” to another agency.

Fourth, ban disclosure of ALPR data for purposes of investigating acts and statuses that are lawful in-state. Three states (Illinois, New Jersey, and New Mexico) forbid disclosure to investigate immigration status. Two states (Illinois and New Mexico) forbid disclosure in investigate healthcare that is lawful in-state. One state (New Mexico) forbids disclosure to investigate protected assembly, petition, and speech.

Fifth, ban any sale of ALPR data. Six states do this: Arkansas, Maryland, Oklahoma, Utah, Virginia, and Washington. California forbids its highway patrol from doing so. New Mexico forbids any disclosure to a private entity (or indeed to any non-law enforcement agency). Data brokers are a menace to privacy and other human rights.

State lawmakers must ensure that the public has access to ample information about the ALPR mass surveillance in their communities. Many people will use this information to resist ALPRs.

First, state laws should require local governments to provide at least 30 days of public notice before spending funds, accepting grants or private donations, or entering into any contract or pilot program for ALPR technology. All proposed vendor contracts and funding sources must be publicly disclosed prior to approval. This helps local officials to exercise their procurement power to reject ALPRs.

Second, laws should mandate the creation of a centralized, public-facing state registry of all police agencies using ALPRs, and the specific locations of all police ALPRs. Sources like DeFlock and EFF’s Atlas of Surveillance include crowd-sourced maps of ALPR technologies, but they are not exhaustive. This was one of the provisions in Minnesota Minn. Stat. § 13.824, which in part, requires police agencies to notify the state within 10 days of installing any ALPR, and requires a state agency to aggregate this information and publish a master list of the locations of all ALPRs (unless a judge determines that disclosure of a particular ALPR’s location poses a genuine security risk).

Third, mandate regular, independent, public audits of agency ALPR use, access logs, and query justifications to ensure compliance with all applicable laws and policies. Five states (Idaho, Minnesota, Montana, Virginia, and Washington) require regular audits, ranging in frequency from monthly to biennial. Three states (Idaho, Maryland, and Washington) require audits of access logs. Minnesota requires an independent auditor and public access to the results.

Fourth, laws must require each agency operating an ALPR system to publish regular reports on its website detailing key operational metrics. Six states require such reports annually: Connecticut, Maryland, New Hampshire, New Mexico, and Vermont. Arkansas requires them twice per year, and Oregon requires them monthly. These states variously require reporting of: the number of ALPRs, and a list of the locations of fixed ALPRs; the number of hotlists used, and a list of them; the number of plates scanned; the number of apparent matches to hotlists, and how many were actual matches; the number plate reads stored; and the number of searches of stored data, broken down by the purpose and agency of search.

Finally, any new ALPR legislation must maintain a proper balance between transparency and privacy in the use of public record acts (PRAs) to obtain ALPR information. Public records and public scrutiny of ALPR programs have shown that people are harmed by these systems. So lawmakers should not cut off access to public records that document these abuses. For example, the public should have access to operational policies, audit logs, vendor communications, and aggregate statistics. EFF opposes blanket exemptions from PRAs that block public oversight of APRs—such as recent bills in states like Arizona. The better approach to protecting individual location privacy is a flexible, case-by-case balancing test already provided by many PRAs. Also, when a document contains both exempt and non-exempt information—like an audit log that contains both motorists’ plate numbers and officers’ search reasons—open records laws should require disclosure of the latter while withholding of the former.

State lawmakers should adopt three additional bans to curtail the footprint of police ALPR surveillance.

First, ban the placement of ALPRs or the collection of ALPR data near healthcare facilities, immigration services, schools, places of worship, courthouses, food banks, and other sensitive locations. Washington and Connecticut do this.

Second, ban police from purchasing ALPR data. Utah and Washington do so. Also, prohibit police from acquiring ALPR data from a private entity, except by means of a warrant. Washington does so. Utah took a step in this direction, requiring a court order based on reasonable suspicion. Absent this type of rule, police evade limits on seizing data from third parties by simply purchasing it from brokers.

Third, ban police and their vendors from using ALPRs or stored ALPR data to power predictive policing technology. EFF opposes such technology and urges state legislatures to ban it. ALPR companies have been providing law enforcement clients with an “analytics” feature, which ingests data from across police agencies, and produces a report of alleged crime trends, hotspots, and corridors. State legislators should prohibit this dangerous use of ALPR systems and data.

A law without strong enforcement is just a piece of paper. EFF recommends three kinds of enforcement measures, all of which appear in existing ALPR statutes.

First, all privacy laws need a “private right of action,” meaning ordinary people can sue a police department that breaks the law. An excellent model appears in the federal Wiretap Act. It expressly authorizes a lawsuit by “any person” whose personal information is processed “in violation of” this law. Privacy harm is harm, and so a person should have their day in court, even if they cannot prove a collateral harm, such as physical injury or economic loss. Moreover, the Wiretap Act provides a full arsenal of remedies to a successful plaintiff, including injunctive and declaratory relief, liquidated and actual damages, and attorney fees.

At least three states enforce their ALPR statute with some version of a private right of action. These are good first steps but need improvement. Washington comes closest: it allows suit to enforce all ALPR rules and provides all remedies. Unfortunately, the plaintiff must prove “injury,” which may shut the courthouse doors to some people whose data is unlawfully processed. Connecticut allows suit by a person “aggrieved” by violation of some ALPR rules. But a person should be able to enforce all rules, and should not have to prove they were aggrieved. Oregon allows suit against a police vendor that “intentionally or with gross negligence” uses or discloses ALPR data in an unlawful manner. But this does not allow any suit against police, or even against a vendor that systematically violates the rules through ordinary negligence.

Second, privacy laws applicable to the police commonly ensure enforcement through the “exclusionary rule.” This means that if evidence is gathered in violation a privacy law, it must be excluded from court proceedings. This is sometimes called “suppression” or “inadmissibility.” The federal Wiretap Act has a strong version of this rule, clearly excluding evidence derived from violations of all rules. The so-called “good faith” exception has weakened many exclusionary rules, so ALPR statutes should expressly reject this exception (though some courts have interpreted some statutory exclusionary rules to have no good faith exception).

At least three states enforce their ALPR statute with some version of the exclusionary rule. Again, these are good first steps but need improvement:

• Virginia comes closest: “Any evidence obtained as a result of a violation” of some of the ALPR rules “is not admissible.” A better rule would apply to all ALPR rules.
• Connecticut provides that ALPR data “may be admissible … provided such data was collected or otherwise acquired or held or stored in accordance with the provisions of this section” (or obtained pursuant to warrant). A more clear exclusionary rule would “exclude evidence unlawfully processed,” as opposed to “admit evidence lawfully processed.” The exclusionary rule should also extend to data unlawfully used or disclosed (and not just to data unlawfully collected or stored).
• Washington provides that information obtained from “a knowing violation” of some of the statute’s rules is “inadmissible.” But the exclusionary rule should apply to all substantive limits (not just some), and to all violations (not just knowing ones).

Third, police agencies and their private contractors that recurringly violate ALPR rules should be forbidden from using ALPRs at all. In New Hampshire, if an agency “repeated[ly]” violates the ALPR statute, the state department of justice may “seek an injunction banning [its] use” of ALPRs. In New Mexico, if a vendor “fails to comply with” that state’s ALPR law, it “may be subject to … suspension of authorization or termination of access.”

Some states impose workplace discipline or criminal penalties on officers who violate ALPR laws, including Connecticut, Idaho, and Missouri. This is no substitute for the three types of strong enforcement measures above. ALPRs are a systemic problem requiring institution-wide enforcement of legal constraints, and not just episodic sanctions against individual officers.

Mass surveillance ALPRs fundamentally threaten digital rights, making an outright statewide ban on police ALPRs the only effective way to fully protect privacy and safety.  The next best thing a legislature can do is limit ALPRs to real-time comparison of passing cars to hotlists, by banning storage of ALPR data beyond a few minutes. If a legislature allows storage, it must impose a tight deletion deadline (days and not months), and strict limits on search and disclosure. A legislature must also ensure transparency and enforcement. 

For convenience here is a concise list of EFF’s ALPR state legislative goals.

SAN FRANCISCO – The Electronic Frontier Foundation (EFF) has published “Online Privacy for Parents and Youth,” a new guide to help families protect themselves online. 

"Parents don’t need to be technology experts to help their children navigate the digital world safely” said EFF Public Interest Technology Director Babette Ngene. “At EFF, we believe in empowering families with the knowledge and confidence to make informed decisions about their digital lives. It’s about meeting families where they are and giving parents the guidance they need to have meaningful conversations with their children about digital safety, based on what works best for their families.” 

EFF launches this crucial resource as state and local governments consider enacting well-meaning but ill-advised laws that cut young people off from essential information and experiences, particularly harming vulnerable youth and marginalized groups who often find safety in supportive online communities they can't access offline. Such laws also entail age gating, which compromises the privacy of all internet users and stifles young users’ ability to learn, grow, and express themselves. 

“Deciding how to help your young people use the internet—including engaging with video games, social media, web search, and everything else—is a complicated topic that touches on many different skillsets and ideas,” said EFF Senior Security and Privacy Activist Thorin Klosowski. “But ultimately it's up to families – not governments – to decide when and how to engage with these sorts of questions.” 

Each young person has unique needs, and, just like their wardrobe, that changes as they age. Parents can engage in thoughtful planning to find the best, most efficient solutions that evolve along with their children.  

EFF’s guide is about helping parents explore their concerns and identify the threats their children face – like trusting AI chatbots or sharing personal information with strangers – and then determine steps they and their children can take to mitigate those risks while still using the games, educational materials, and communication platforms that help them grow. 

The guide also compiles links to parental control setup documentation for a variety of popular devices and platforms, and advises parents on quick, easy tasks they can perform to advance their children’s online privacy. 

For the guide: https://ssd.eff.org/module/online-privacy-for-parents-and-youth

In a victory for data privacy, a federal court in New York definitively ruled that the Office of Personnel Management (OPM) violated the law when it disclosed millions of federal employee records to DOGE agents. In a summary judgment ruling issued on October 20, 2026, the court found that OPM violated two provisions of the Privacy Act—one restricting disclosure of personal records between agencies and another requiring agencies to protect the security and confidentiality of records.

Since the case was filed, many DOGE agents have left the government and OPM has rescinded their database access. However, OPM should still clean up the cybersecurity mess it created. The court will decide the remedy at a later date.

"By virtue of the Court’s summary judgment ruling, OPM’s decision to grant DOGE Agents broad access to its systems of recordings containing [personally identifiable information (PII)] has been held unlawful," Judge Denise Cote wrote.

"The record also provides no support for a finding that it was the pursuit of a modernization agenda that prompted OPM to grant sweeping and expedited access to a score of DOGE Agents to its systems of records containing PII," she added.

OPM maintains databases of highly sensitive personal information about tens of millions of federal employees, retirees, and job applicants. Starting in January 2025, OPM disclosed data to the U.S. DOGE Service, initially led by Elon Musk. This disclosure violates the federal Privacy Act of 1974, a watershed anti-surveillance statute that prevents the federal government from abusing our personal information.

EFF sued OPM and DOGE because this data disclosure violates the Privacy Act. We represent unions of federal employees – the American Federation of Government Employees and the Association of Administrative Law Judges – and individual employees. We seek injunctive and declaratory relief. We filed in the U.S. District Court for the Southern District of New York.

Our co-counsel are Lex Lumina LLP, State Democracy Defenders Fund, and The Chandra Law Firm LLC.

Donate to EFF

Support Privacy & Government Accountability

This post was co-written by Andi Cao, an EFF legal intern.

Automated license plate readers (ALPRs) are mass surveillance. They photograph every passing car and log where and when it was captured, building a searchable record of where nearly everyone drives. Cities should not buy them, and cities that already have them should shut them down.

Where these systems still operate, courts and state legislatures must reduce their harms by (among other things) requiring police to get a warrant before searching stored ALPR data. A warrant requirement would not make mass surveillance acceptable, but it would require police to justify a search to a judge before accessing records of people’s movements. Without this independent judicial check on police power, officers have been searching this data for improper reasons (many exposed by EFF) that violate our fundamental rights.

A warrant requires a judge to find probable cause and specificity—many warrantless police searches of ALPR data lack both. Officers have searched this data to profile Romani people, monitor protesters, investigate an abortion, find immigrants for deportation, track a man who complained about a deputy, stalk romantic partners, and even for no stated reason beyond “LMAO.” A warrant requirement would force law enforcement to justify these searches before accessing sensitive location records, not leave the people they target to discover and challenge the abuse after the harm is done.

Warrants are a check on police power. Before officers can search a person or their property, the Fourth Amendment requires them to convince a neutral judge, under oath, that probable cause exists, meaning a fair probability the search will turn up evidence of a crime. The warrant must also describe with particularity what will be searched, so police cannot go on a fishing expedition. Critically, the officer who wants to search is not the person who decides whether the search is justified.

That is why warrants are required for so many kinds of police surveillance. For example, police generally need a warrant to wiretap phone calls, to search the contents of a cell phone, and to obtain historical cell phone location records.

Stored ALPR data reveals sensitive information about the private details of our lives: where we go, when, and with whom. That’s why we believe police searches of stored ALPR data should also require a warrant. Earlier this month, a federal judge in Oklahoma agreed.

San Jose demonstrates the harmful consequences of letting law enforcement search stored ALPR data without a warrant. EFF and the ACLU of Northern California represent SIREN and CAIR-California in a lawsuit challenging this practice.

The first problem is the sheer number of times that SJPD officers search stored ALPR data: 261,711 times between June 5, 2024, and June 17, 2025, about 692 searches a day. That scale is the result of having no warrant requirement, which would greatly reduce this kind of casual police surveillance.

Inevitably, officers abuse this unchecked power. In September, San Jose’s police chief disclosed an example from five months earlier: the department had fired an officer who used its ALPR system to track a woman who had accused his cousin of domestic violence, and then shared her location information with his cousin. The abusive search of stored ALPR data came to light only after the woman told police that her alleged assailant had implied he was tracking her.  A warrant requirement would have prevented this.

Of course, similar problems appear nationwide. For each example below, consider how things could have turned out differently if the officer had to put the reason for the search in a sworn warrant application and handed it to a judge.

More than 80 law enforcement agencies searched the Flock network using terms that unfairly stereotype Romani people, often without mentioning any crime. For example, the Grand Prairie (Texas) Police Department searched for “g*psy” six times using a Flock feature that identifies vehicles traveling together, in effect targeting a whole community without naming a crime.

No judge should sign that warrant. Ethnicity is not evidence of a crime, and a search aimed at an entire ethnic group is the opposite of particular: it targets people for who they are, not evidence that they committed a crime.

EFF found that more than 50 agencies ran hundreds of protest-related searches of ALPR data in 2025. For example, the Tempe (Arizona) Police Department ran three searches tied to the No Kings protest across 425 networks, with no identified crime.

This lacks probable cause and targets activity that is protected by the First Amendment.

In 2025, a Texas sheriff’s office searched more than 83,000 Flock cameras, including cameras in Washington and Illinois, where abortion is legal, listing the reason as “had an abortion, search for female.” Records later obtained by EFF show deputies had opened a “death investigation” after the woman’s partner, who was later charged in connection with assaulting her, reported her self-managed abortion.

In a state where abortion is legal, a judge would not grant a warrant for a search like this.

San Francisco police let out-of-state agencies search the city’s ALPR data (which itself violates California law), and at least 19 of those searches were marked as related to ICE (though San Francisco is a sanctuary city). This problem is widespread. One investigation found more than 4,000 cases in which police departments around the country searched stored ALPR data on behalf of federal law enforcement, including for immigration investigations. Some agencies even connect their ALPR systems to an “Immigration Violator” hotlist populated by ICE, which includes plates tied to administrative warrants issued by ICE agents without any judge involved.

Here too, a judge in a state where police may not help enforce immigration law would not issue a warrant for these searches.

In May 2025, Navy veteran Napoleon Jones lawfully recorded a police traffic stop of another person in Sussex, Wisconsin. Jones alleges that a Waukesha County deputy retaliated by stopping and arresting him. He was released five hours later without citations or charges, and the deputy’s supervisor concluded the stop was illegal. Jones filed a complaint against the deputy the next day.

In the weeks that followed, officers reportedly ran more than 100 Flock searches against Jones, including some run by the deputy he had complained about. Deputies testified that a lieutenant ordered the searches in connection with Jones’s complaint.

No judge should authorize tracking a man for criticizing the police.

The Institute for Justice (IJ) has documented dozens of cases in which officers allegedly used ALPRs to track current, potential, and former romantic partners. For example, a Lufkin, Texas, officer was indicted on 100 felony counts after allegedly searching for his ex-girlfriend and people she knew more than 10,000 times. Likewise, a Milwaukee officer allegedly tracked a woman he was dating and her ex nearly 180 times over two months. IJ found that most of these cases surfaced only after victims reported the officer.

We hope that no officer would write “tracking the location of my ex-girlfriend” on a warrant application, and even if they did so, a judge would not authorize such a search.

Sometimes officers do not even provide a real reason to search stored ALPR data. EFF found officers logging search reasons such as “LOL,” “LMAO,” “idk,” and random keyboard mashing. These entries are supposed to document why a search complies with an agency’s own use policies. A Goshen (Indiana) Police Department officer searched 6,474 networks, representing data from 82,413 cameras, with “idk” as the reason.

An officer who asked a judge for a warrant to track someone for “LMAO” would be laughed out of the courtroom.

The best way to eliminate these abusive searches is to end ALPR mass surveillance. Where these systems remain, however, courts and state legislatures must require police to get a warrant before searching stored ALPR data. Warrants are the time-tested way to put an independent judicial check on police use of a dangerous surveillance technology. Officers must establish probable cause and specifically describe the proposed search before accessing the data, while existing exceptions for genuine emergencies still apply.

These protections must also cover real-time tracking. An ALPR “hotlist” is a list of license plates that trigger an alert when a camera spots a matching plate. Officers can stalk potential romantic partners by putting them on a hotlist, as a Florida deputy allegedly did with a woman he met on the set of a TV show. Thus, state legislators should require officers to get a warrant before putting a person on an ALPR hotlist (as well as a warrant to search stored ALPR data).

Exciting news for Safari users: Privacy Badger is now available on Safari for macOS! You’ll need Safari version 26.5 or higher to use it.

Whether you want to protect yourself from data brokers or just don’t want Big Tech profiting off your browsing data, Privacy Badger is here to help. More than 4 million people already use EFF’s free tracker-blocking extension to take a stand against online surveillance. Safari users can now join them and EFF in the fight for a better web.  

Install Privacy Badger

Support the Fight Against Online Tracking

Safari support has been one of Privacy Badger’s most requested features for years. Thanks to recent improvements in Safari's extension support, we’ve finally made it happen for Safari version 26.5 and higher. Privacy Badger is now available on all major desktop browsers. And before you ask, yes, we will look into supporting Safari for iOS once we've ironed out any issues on macOS.

Most websites contain invisible tracking code that lets companies collect, share, and profit from your personal data. Some of those companies are data brokers that sell your information to anyone willing to pay. Creepy ads are just the tip of the iceberg when it comes to the harms of online tracking: Once your information enters the data broker ecosystem, it’s nearly impossible to know who buys it or what they’re doing with it. Privacy Badger blocks online trackers to prevent your browsing data from being used against you.  

Privacy Badger goes beyond Safari’s built-in privacy features to provide additional protection against online tracking. In its default browsing mode, Safari limits the information that trackers can access but still allows them to load. Privacy Badger blocks trackers from loading in the first place, preventing them from tracking you using techniques that bypass Safari’s protections. 

Privacy Badger works well by itself, and will continue to enhance your privacy when installed along side an ad blocker or with Safari’s Advanced settings to block trackers in all windows. Since Privacy Badger automatically learns to block trackers, Privacy Badger will block trackers that other tools don’t know about or choose not to block. 

For example, Privacy Badger blocks tracking from widgets like embedded tweets, comments sections, and video players. These widgets can share your data even if you never click on them. Privacy Badger prevents these trackers from loading and replaces them with click-to-activate placeholders. 

Beyond tracker blocking, Privacy Badger automatically sends the Global Privacy Control (GPC) signal to tell websites not to sell or share your personal information. The GPC signal is already legally binding in a growing number of states, including California, Colorado, New Jersey, and Connecticut.  

All these features (and more) work together to help keep your browsing activity private on Safari. 

If you use Safari on macOS, give Privacy Badger a try. You’ll install Privacy Badger through the App Store on your macOS device. Installing extensions is a little more complicated on Safari than other browsers, so follow Apple’s directions here if you’re having trouble.  

This is our first release of Privacy Badger for Safari, and we will keep improving it based on your feedback. If you run into any problems, let us know by emailing [email protected], or by opening an issue on GitHub. And if blocking harmful trackers causes issues on a website you visit, you can disable Privacy Badger for that specific site while keeping it enabled and protecting you everywhere else. You can also let us know about broken sites by using the “Report broken site” button found next to the “Disable for this site” button in Privacy Badger’s popup window. 

When you use Privacy Badger, you join EFF and millions of other users in the fight for a better web. Help spread the word by telling your friends and family who use Safari that they can use Privacy Badger too!  

Install Privacy Badger 

Support the Fight Against Online Tracking 

SAN FRANCISCO—Privacy Badger, the award-winning tracker blocker made by the member-supported digital rights nonprofit Electronic Frontier Foundation (EFF), is now available for Safari on macOS.  

Safari users can download the free browser extension today and join the more than 4 million people using Privacy Badger to take a stand against data brokers and online surveillance. Users will need Safari version 26.5 or higher to use it.

Ranked as one of the best ad tracker blockers by The Wirecutter, Consumer Reports, PC Mag and more, Privacy Badger is different from other blocking extensions. Created by EFF to stop companies from spying on you online, Privacy Badger automatically learns to block hidden trackers on the web, along with additional privacy-protecting features. 

"Most websites contain invisible tracking code that lets companies collect, share, and profit from your personal data," said Lena Cohen, Staff Technologist at EFF. "Creepy ads are just the tip of the iceberg when it comes to the harms of online tracking: Once your information enters the data broker ecosystem, it’s nearly impossible to know who buys it or what they’re doing with it. Safari support has long been one of Privacy Badger’s most requested features and EFF is excited to be able to make this important tool available to Safari users to fight back for a better web." 

On Safari, Privacy Badger goes beyond the browser’s built-in privacy features to provide additional protection against online tracking. In its default browsing mode, Safari limits the information that trackers can access, but it still allows them to load. Privacy Badger blocks online trackers from loading in the first place, preventing them from tracking you using techniques that bypass Safari’s protections. 

With today's launch, Privacy Badger is now available on all major desktop browsers. 

As the leading nonprofit defending online civil liberties, EFF is committed to fighting mass surveillance in the courts, through activism, and by making free, privacy-protecting tools like Privacy Badger. Now, even more people can join the movement against online tracking, by installing Privacy Badger themselves, telling their friends and family who use Safari about it, or by supporting Privacy Badger today. 

For more on Privacy Badger for Safari on macOS: https://www.eff.org/deeplinks/2026/10/announcing-privacy-badger-safari-macos

For more on the dangers of online tracking: https://www.eff.org/deeplinks/2025/03/online-tracking-out-control-privacy-badger-can-help-you-fight-back 

For more on how Privacy Badger works: https://privacybadger.org/ 

Contact: [email protected]

Barry Steinhardt, a lifelong privacy champion and an early leader in the battle for digital civil liberties who served as EFF’s president in 1998-99, died last week at age 73. 

A Connecticut native, Steinhardt led the Vermont and Pennsylvania affiliates of the American Civil Liberties Union before becoming associate director of the national ACLU. There, he helped spearhead the fight to overturn much of the Communications Decency Act, resulting in the U.S. Supreme Court’s landmark Reno v. ACLU ruling in 1997. 

When the internet was still young, Steinhardt became one of the first attorneys to specialize in the intersection of technology and privacy. In early 1998 he took a leave of absence from the ACLU to become EFF’s president, and then returned to the ACLU in 1999. 

In 2002 he founded the ACLU’s Program on Technology and Liberty. After retiring from the ACLU, he continued on as a senior advisor to London-based Privacy International. 

Steinhardt chaired the 2003 Computer Freedom and Privacy Conference and served on a wide variety of significant panels and boards, including the Department of Transportation’s Negotiated Rule Making on national driver’s license standards, the Advisory Committee to the US Census, and the Blue Ribbon Panel on Genetics of the National Conference of State Legislatures. He also was selected to be a member of the US delegation to the G-8 Government and Private Sector Tokyo conference on Cyber Crime and served as an advisor to the Czech Helsinki Committee.  

“Barry led many critical technology and civil liberties fights and was an important mentor to me and others in the field,” said EFF Executive Director Nicole Ozer. “Barry’s influential work across strategy to defend and advance privacy rights and safeguard people against surveillance has made its mark on the law and policy of our digital world and protected the rights and safety of people across the United States and around the world.”