Slop Happens

Flock is trying, and succeeding, to become Google for cops.

Flock Safety runs more than 100,000 license plate readers across the US. A small-town police department can rent them by subscription, then search the entire national network from a laptop: a license plate, a black SUV, even a protest sign.

But reading license plates isn't new. Toll booths do it. Speed cameras do it. So where does a useful tool turn into something else?

In Episode 8, Claudio Luís Vera and Bianca Prins compare the American approach with the Dutch one. In the Netherlands, many traffic cameras only capture an image when someone breaks the law, footage can't be kept indefinitely, and access depends on your role and your reason. In the US, the hard questions often get asked after the product is already on the street.

We talk about what happens when a single-purpose camera gets swapped out for one that can do far more, and nobody votes on the change. We ask who gets to search a national database, what training they have, and who pays the price when the system is misused. Flock has since cut its recommended default data retention from 30 days to 7. Is that enough? Or is it a patch on a question that should have been answered before launch?

Bianca's answer: treat ethics the way we treat accessibility. Build it in at the start.

Ethics after launch is just damage control.

Meta donated 15,000 smart glasses to a disability organization in Ireland.

Free devices. Independence for blind and low-vision users. Good press.

Here's the catch: to use the AI features, you sign off on your privacy. German researchers found the recording indicator doesn't always show. And even ICE — an agency not exactly known for privacy scruples — passed on adopting them.

So is this a genuine accessibility win, or a fig leaf for a much bigger data problem?

In the new episode of Slop Happens, Bianca and Claudio dig into what happens when a human right (accessibility) gets used to sidestep another human right (protection from abuse) — and why the price of an assistive device might be a better predictor of its safety than its marketing.

Note: after this episode was recorded, Meta began running ads in the US announcing that it's donating Meta glasses to disabled vets.

In a past life, Claudio ran a web design boutique that always had the Magic 8-Ball on the conference table, consulted with total mock-seriousness whenever a decision got stuck.

Turns out that's not a bad starting point for evaluating predictive AI. So — are today's algorithms that much better than consulting the Magic 8-Ball?

Not by as much as you'd think. A system running at 76% accuracy is only 26 points better than a coin flip. And that gap is exactly where things get real: it's the difference between who gets released before trial, who gets flagged as a hiring risk, who gets classified as a legitimate military target — and, in Allegheny County, Pennsylvania, which parents got to keep their kids.

This week on Slop Happens, we also dig into an angle that doesn't get enough airtime: hiring algorithms that seek clones of a company's best performers will lock out disabled candidates, and produce different (and differently biased) results depending on whether the training data comes from the US or the EU.

When the algorithm gets it wrong, no one ever gets fired for keeping silent. Which is why checks and balances on faulty AI are so important.

For blind and low-vision users, smart glasses that can read a letter aloud or identify a package on sight are nothing short of transformative. But those benefits rely on a camera that’s always on, always recording, and always sending data somewhere.

Combined with facial recognition, though, they can provide far too much information to a stalker or form a grid of real time facial-recognition for a surveillance state. Same hardware. Same camera. Wildly different outcomes depending on who's wearing it and who's standing nearby.

In this episode, Claudio and Bianca trace what happens when an assistive technology becomes a data pipeline — for facial recognition, for corporate data-selling, for law enforcement. They dig into a leaked Meta feature that quietly built in facial recognition, and photo of a federal agent wearing the glasses at an immigration raid, and explore the risk when someone else's glasses are recording you without your consent.

Is this freedom or surveillance? Who's accountable when the tool works exactly as designed and someone still gets hurt? And is there a way to flag "this device is recording" that doesn't stigmatize the person wearing it?

Florida, USA meets The Netherlands for a conversation that doesn't land on easy answers.

Mentioned in this episode:

• WIRED — Meta Silently Added Face-Recognition Code for Its Smart Glasses to Millions of Phones
• Futurism — Meta Had the Worst Possible Response When Its Workers Were Watching Naked Footage of Its Ray-Ban AI Glasses Users
• 404 Media — A CBP Agent Wore Meta Smart Glasses to an Immigration Raid in Los Angeles

Correction: In this episode, the agent photographed wearing Meta smart glasses at a Los Angeles immigration raid is referred to as an ICE agent. Per 404 Media's reporting, the agent was with Customs and Border Protection (CBP), a related but separate agency.

Your ballot is secret. Your donation history isn't. Your voter registration isn't. And in the age of algorithmic profiling, the data trail you leave before you vote may matter more than the vote itself.

In this episode of Slop Happens, Bianca and Claudio dig into something that doesn't get enough attention: not whether elections get hacked on election night, but how the data surrounding them gets weaponized long before anyone casts a ballot.

We walk through how Dutch paper balloting actually works — volunteers, sealed packages, signed-off destruction orders — and why that transparency is the feature, not the bug. We revisit the Tascón List (from Episode 3) and what it tells us about petition signatures, donor records, and the permanent data trails that political acts leave behind.

The uncomfortable question: if you signed a petition today, would you be comfortable with it surfacing in a background check ten years from now?

Paper isn't a primitive fallback. It's tamper-proof by design.

Paper, trust, and democracy — and why all three are harder than they look.

In the age of AI, it’s easy to forget that it doesn’t take sophisticated tech to build an instrument of power. In the 2000s, Venezuela gave the world a masterclass on how a simple, flat database could turn a petition into a tool of repression.

In this episode of Slop Happens, we unravel the dark legacy of the Tascón List —a petition for a recall referendum that turned into a political blacklist that targeted the opposition. Fast-forward to today: could the misuse of personal data, voting records, or social profiles create modern versions of these lists? We break down the risks, safeguards, and what we must do to protect our society from history repeating itself.

Is the same dynamic playing out today, with government funding and contracts being tied to political loyalty? Is there a risk to signing a public petition, if there’s no guarantee that the data ever really disappears?

Bianca and Claudio talk about what it takes to weaponize civic data, how the US voter registration system already exposes your party affiliation, and why signing a public petition today is a decision you might be living with for decades.

New technologies, age-old tactics.

In this second episode of Slop Happens, Bianca Prins and Claudio Luís Vera welcome Maranke Wierenga, advisor on data-driven work and member of the NEN Commission on AI & Big Data, where she also chairs the Working Group on Ethics and Fundamental Rights.

Together, they explore curious cases where technology fails the people it’s meant to serve.

First, they dive into the COMPAS case in Florida, where a risk-scoring system used in the justice system showed troubling bias across racial groups. The discussion looks at how “neutral” algorithms can reinforce inequalities already present in society.

They then examine a case from the Netherlands, where automated parking enforcement unintentionally creates barriers for people with disabilities.

What happens when efficient systems overlook real people? And how can we design technology that is fairer and more accountable?

Join us for this conversation—and let us know what you think.

Claudio and Bianca welcome you to their new podcast, Slop Happens, and discuss the weirdness that happens when technology goes wrong.

One evening, Claudio sees a trippy cloud with shining debris overhead. The next day, he discovers it was a SpaceX mission that exploded and rained debris down over Florida, the Bahamas, and the Turks and Caicos Islands.

SpaceX issues a tweet on X, treating the explosion as a teachable moment that offers lessons for long-term success. But as if that weren't weird enough, the debris is actually SpaceX property – even if it comes hurtling through your roof.