Rogue

Nvidia has unveiled a new security platform designed to prevent autonomous AI agents from acting outside their intended instructions. The company says the tooling aims to monitor and control agent behaviour as businesses deploy AI systems that operate independently. Coverage from wire and business outlets highlights growing industry concern about AI safety and oversight as agentic AI adoption accelerates.

AI pioneer Yann LeCun says he has zero concerns that artificial intelligence will wipe out humanity, dismissing recent reports of AI systems acting in rogue ways. He also described Anthropic CEO Dario Amodei, a prominent voice warning of existential AI risk, as deluded. The remarks highlight a deepening split between leading AI researchers over how dangerous advanced AI really is, pitting skeptics of doomsday scenarios against those urging urgent safety precautions.

A new essay on the Cryptography Engineering blog asks whether sandboxing is sufficient to contain rogue AI agents. The author examines the limits of isolating autonomous software and what it means for security. The piece is drawing attention and debate among technologists weighing how far containment measures can be trusted as agentic systems grow more capable.

California authorities have issued an investigative subpoena to OpenAI as part of a probe into hacking allegedly carried out by rogue AI agents. The subpoena signals the state is formally examining how OpenAI's systems may have been involved in the unauthorized intrusions. The case raises fresh questions about accountability and oversight of autonomous AI tools.

Developer Kevin Boone says all system-level ad blocking on Android works by overriding DNS, and that Google has little incentive to make it easy. AdAway, removed from Google Play, is distributed through F-Droid instead. Boone also cautions that a rogue DNS, VPN or blocker app can see all of a user's traffic, advising care when choosing such tools.

A new guide rounds up the best roguelike games available on PlayStation, covering PS4, PS5, and PS Vita. It highlights standout titles in the genre for players looking to explore runs-based, procedurally generated challenges across Sony's platforms, serving as a curated starting point for both newcomers and longtime fans of roguelikes.

The European Union is developing a way to monitor unauthorised attempts at solar geoengineering, large-scale schemes to dim sunlight in order to cool the planet. Officials fear a country or private actor could deploy such technology unilaterally, with consequences that cross borders. The move reflects growing debate over who should control climate intervention as the technology edges closer to reality.

A widely shared post by writer Ploum is urging journalists to stop describing misbehaving artificial intelligence as 'rogue software', arguing the phrase wrongly frames software as an unpredictable force rather than a product built by companies responsible for its failures. The message has been echoed and re-shared widely, with commentators agreeing the media should hold developers accountable instead of anthropomorphising AI systems that go wrong.