Cyber Ethos

Cyber Ethos

By Ben HermannTechnology
Download on the App Store

Cyber Ethos episodes

  • AI Transparency Becomes Law: Finland, France, and the FTC Respond

    Send us Fan Mail

    Three regulators, one week, and the first real test of AI Act transparency

    This week, three regulators moved on AI transparency and pricing fairness within days of each other.

    Finland's Traficom activated supervision of most EU AI Act transparency duties on 20 August, covering direct AI interaction disclosure, machine detectable synthetic content and mandatory biometric notice. Most duties have applied since 2 August 2026, with a transition for legacy system marking running to 2 December 2026.

    France's CNIL updated its AI Act guidance on 17 August to confirm GDPR obligations run independently of AI Act compliance, even where documentation overlaps, following the AI Omnibus that entered into force on 27 July.

    In the United States, the FTC opened consultation on 19 August on a policy statement that would require disclosure whenever a price is personalized using personal data, rather than banning the practice outright. Comments close 18 September 2026.

    Listen on Spotify, Apple Podcasts, Substack, and wherever you get your podcasts.

    Check out Cyber Ethos on cyberethos.substack.com (English) or cyberethosde.substack.com (Deutsch)

    8 min
  • S1E20: Can You Spot the AI?

    Send us Fan Mail

    Inside the Industry's New Transparency Push

    This week Anthropic, Spotify, Suno and Substack all rolled out new AI-content labelling. Anthropic now automatically watermarks everything Claude generates, citing EU AI Act Article 50. Spotify is badging AI artist profiles. Suno added watermarking and fingerprinting. Substack built in the AI detector Pangram. A Deezer study found 97 percent of listeners cannot tell AI-generated music from the real thing. We look at whether watermarking actually holds up, and what happens once someone builds a way around it.

    Check out Cyber Ethos on cyberethos.substack.com (English) or cyberethosde.substack.com (Deutsch)

    6 min
  • S1E19: When Your Car Reports You

    Send us Fan Mail

    The EU's Plan to Auto-Limit Speed

    A new EU proposal could let your car's engine decide when you've gone too fast.
     
     This week on Cyber Ethos, we look at the European Commission's early-stage investigation into a satellite- and camera-based system that would automatically restrict engine power when a vehicle exceeds the speed limit, no warning, no easy override. We unpack why this shifts a well-understood advisory system, ISA, into a safety-critical actuator with a much larger attack surface, and what that means for drivers, regulators and anyone thinking about the security of connected vehicles.
     
     Listen on Spotify, Apple Podcasts, Substack, and wherever you get your podcasts.

    Check out Cyber Ethos on cyberethos.substack.com (English) or cyberethosde.substack.com (Deutsch)

    7 min
  • AI Regulation: Delays, Hurdles, and NeEurope Rewrites Its AI Rules: Delayed Deadlines, Tougher Bansw Prohibitions

    Send us Fan Mail

     | S1E18 – The EU’s Digital Omnibus delays high-risk AI deadlines but adds new bans and mandatory marking of all synthetic content
     | The EU’s Digital Omnibus on AI entered into force on 27 July 2026, rewriting the AI Act’s implementation timeline. High-risk AI deadlines are pushed to December 2027, but new prohibitions on CSAM and non-consensual intimate imagery, plus mandatory machine-readable marking of all synthetic outputs, make the rules tougher where it matters most.  This episode covers: the 16-month deadline extension for high-risk systems. New prohibited practices not in the original Act. Transparency obligations starting 2 August 2026. Italy’s enforcement against Character Technologies over minors. EDPB guidelines on web scraping for generative AI. The stakeholder consultation risk for democratic governance.  Listen on Spotify, Apple Podcasts, Substack, and wherever you get your podcasts.

    Check out Cyber Ethos on cyberethos.substack.com (English) or cyberethosde.substack.com (Deutsch)

    8 min
  • Scanned First, Regulated Later: Western Australia's Live Facial Recognition Trial

    Send us Fan Mail

    S1E17 - Western Australia's live facial recognition trial, the real error rate behind the headline figures, and the privacy law that arrived nine days too late

    Western Australia Police became the first police force in Australia to deploy live facial recognition in public spaces. In five weeks, the system scanned more than 131,000 faces, generated 33 alerts, and led to 19 arrests - all before the state's new privacy law had fully come into force, and without the state's own privacy watchdog being consulted on the trial's design.

    This episode covers: how live facial recognition differs from a normal camera, the real false-alert rate hiding behind the headline scan count, the regulatory gap that let the trial begin before the law caught up, and why Australia has no equivalent to the EU's restrictions on real-time biometric surveillance.

    Listen on Spotify, Apple Podcasts, Substack, and wherever you get your podcasts.

    Check out Cyber Ethos on cyberethos.substack.com (English) or cyberethosde.substack.com (Deutsch)

    8 min
  • How Old Are You, Really? Inside the EU's Privacy-First Bet on Child Safety

    Send us Fan Mail

    S1E16 – Inside the EU's zero-knowledge push to verify age online without collecting identity
    The EU's new age verification app promises to confirm you're old enough for a platform without ever learning who you are. This episode unpacks the Commission's Recommendation, the DSA enforcement actions against Meta and Snapchat that made it necessary, the zero-knowledge proof design behind it, and its surprising new reach into AI companions and chatbots.  This episode covers: the enforcement gap in self-declared ages; zero-knowledge proof verification and the EU Digital Identity Wallet; the patchwork of national minimum ages; AI chatbots and companions entering the scope; and the circumvention risks the Commission openly acknowledges.  Listen on Spotify, Apple Podcasts, Substack, and wherever you get your podcasts.

    Check out Cyber Ethos on cyberethos.substack.com (English) or cyberethosde.substack.com (Deutsch)

    8 min
  • The Companion Trap: Why People Choose General AI Over Purpose-Built AI for Mental Health

    Send us Fan Mail

    Why general-purpose AI keeps winning the mental health advice fight, and what gets lost in the process
    Hundreds of millions of people already ask general-purpose AI like ChatGPT and Claude for mental health advice, not because it was built for that, but because it is already open on their screen. A new analysis from AI scientist Lance Eliot breaks down why purpose-built mental health AI keeps losing to pure convenience.  This episode covers: the structural reasons general-purpose AI wins by default, the five patterns of how people actually split their usage, the risk of losing context when switching tools mid-crisis, and why the commercial incentives make this gap unlikely to close on its own.  

    Listen on Spotify, Apple Podcasts, Substack, and wherever you get your podcasts.

    Check out Cyber Ethos on cyberethos.substack.com (English) or cyberethosde.substack.com (Deutsch)

    5 min
  • When the Algorithm Votes: AI, Democracy, and the Fight for Digital Freedom

    Send us Fan Mail

    How AI has embedded itself inside democratic systems, and what it will take to keep citizens in control.

     | AI already curates the news you read, profiles voters at scale, moderates online debate, and assists governments in allocating resources. A landmark 2026 handbook from 59 researchers across 12 countries frames the defining question: who controls AI in your democracy – and on whose terms?  This episode covers: voter surveillance, profiling and targeting; privacy as democratic infrastructure; misinformation, deepfakes and algorithmic polarisation; the EU AI Act's democratic limitations; and democratic resilience by design.  Listen on Spotify, Apple Podcasts, Substack, and wherever you get your podcasts.

    Check out Cyber Ethos on cyberethos.substack.com (English) or cyberethosde.substack.com (Deutsch)

    8 min
  • Scan Everything, Say Nothing: Germany's Chat Control Deception

    Send us Fan Mail

    Germany says it supports targeted measures on private message scanning. Its EU negotiators are pushing for mass surveillance of all private messages in Europe. These are not the same policy.
    This episode covers: what client-side scanning does to end-to-end encryption; why the child safety framing does not change the surveillance architecture; what the EU AI Act disclosure requirements reveal about the limits of transparency-based governance; and what Pope Leo XIV warning about opaque algorithms has to do with democratic accountability in AI systems.
    Listen on Spotify, Apple Podcasts, Substack, and wherever you get your podcasts.

    Check out Cyber Ethos on cyberethos.substack.com (English) or cyberethosde.substack.com (Deutsch)

    7 min
  • Europe's AI Deepfake Ban – A Landmark Step With Limits

    Send us Fan Mail

    The EU Parliament has voted to ban AI applications that generate sexualised deepfakes, amending the European AI Act with explicit prohibitions on non-consensual synthetic intimate imagery.
    In this episode, Cymon and the Cyber Ethos team examine what this legal milestone actually changes: the enforcement challenge on global platforms, the gap between prohibition and victim redress, and the broader EU push for digital sovereignty through Chip Act 2.0 and the Cloud and AI Development Act.
    Key points covered:
    - What the AI Act amendment actually prohibits and why it matters
    - Why enforcement across global platforms is the next hard problem
    - The Digital Omnibus package and its wider implications
    - What the law still leaves unresolved for victims
    - The ethical question: banning a tool vs. preventing the harm
    Listen on Spotify, Apple Podcasts, Substack, and wherever you get your podcasts.

    Check out Cyber Ethos on cyberethos.substack.com (English) or cyberethosde.substack.com (Deutsch)

    8 min

About Cyber Ethos

From the publisher's feed

Weekly insights on online privacy, cybersecurity, and AI ethics. Clear, honest, no jargon. New episodes every Thursday.