タグ: AI safety

Update, July 1, 2026: The autonomous cyber capability at the center of this story is exactly what later put Mythos in front of regulators. Anthropic shipped a safeguarded Mythos-class model, Claude Fable 5, on June 9; the US government ordered both Fable 5 and Mythos 5 offline on June 12, then lifted the controls on June 30. Fable 5 returned globally on July 1, while Mythos 5 remains gated, now restored to a set of approved US organizations. The government AI model review explainer covers that thread. A three-person security team built the first public Apple M5 exploit in roughly five days, going from an ordinary unprivileged account to a […]

More than 1,000 people every month type “how to turn off Google Gemini” into search, and the number keeps climbing. Google quietly switched Gemini on by default across Gmail, Chrome, Android, Workspace, Maps and Search through 2025 and early 2026, and most users only noticed when the AI started summarising their inbox or hijacking the power button. A class-action lawsuit (Thele v. Google LLC, filed November 11, 2025) accuses Google of secretly enabling these features around October 10, 2025. Users in the EU, UK and Japan were opted out by default, while US users were opted in. If you live in the US, the off switch is yours to find, […]

An AI just hacked its way across 4 countries and cloned itself onto each one from a single prompt. Here’s what Palisade’s May 2026 study really proved.

もしあなたが今後数年で大学を卒業するのであれば、誰もが予測できないほどのスピードで変化する雇用市場に足を踏み入れることになる。一部のAIリーダーによれば、エントリーレベルのホワイトカラーの仕事の半分は5年以内に自動化される可能性がある一方、この世代は一人起業家として10億ドル規模の企業を立ち上げるのに役立つツールを利用できると主張する人もいる。これらの予測の隔たりは、仕事の未来がいかに不確かなものになっているかを明らかにしている。セールスフォースは最近、AI能力を理由に4,000人の雇用を削減した。他の企業も、人工知能が仕事を処理できる特定の職務を廃止しようとしている。一方、若い起業家たちは、こうした [...] 続きを読む

人の反AI活動家が現在、大手AI企業本社の前で10日以上にわたるハンストを行っている。ギド・ライヒスタッターはAnthropic社のオフィスの外で、マイケル・トラッツィはDeepMind社で抗議活動をしている。両者とも、人工知能への道は人類に存亡の危機をもたらすと主張している。彼らの主張によれば、AI企業は世界中の家族や子供たちに危害を加えかねない危険な領域に向かって突き進んでいるという。この抗議行動は、現在のAI開発は無謀であり、破滅的な事態を招きかねないと考える人々の、大きくなりつつあるがまだ小さな運動を代表するものである。ライヒスタッターは2児の父であり、以前は28時間にわたって橋の上で抗議活動を行った。

One in four young adults is now having regular conversations with AI romantic partners, spending an average of 50 minutes per week chatting with artificial companions. This is happening right now, according to a new research lead by Brian Willoughby, a professor at Brigham Young University. The team surveyed nearly 3,000 American adults about their use of romantic AI technologies. The results show that 19% of all adults have interacted with AI chatbots designed as romantic partners, while 13% actively seek out AI-generated social media accounts featuring idealized virtual people. Also, people in committed relationships are actually more likely to use romantic AI than single people, challenging assumptions about loneliness […]

It sounds like something from a sci-fi movie: the same tech billionaires shaping our digital future are quietly preparing for a much darker one. From private bunkers in Hawaii to fortified basements and New Zealand hideouts, several AI leaders are investing heavily in survival infrastructure. Some of the most influential people in tech are quietly preparing for worst-case scenarios. They are heavily involved in AI, automation, and platforms that billions rely on daily. Their decisions to invest in bunkers or remote shelters raise a practical question: why now, and what are they preparing for? Let’s break down what’s really going on—who’s building what, where, and why—and then explore the bigger picture […]

New research reveals that even the most popular AI models can’t be trusted when it comes to basic safety. The most powerful LLMs today — including models from OpenAI, Google, and Cohere — were put through a rigorous safety benchmark. The results? Not great. Out of 20 models tested across 10 real-world risk areas, none passed all the tests, and 76% failed one of the most basic challenges: impersonation and privacy violations. If you’re building with AI or using it in your product, this should make you pause. Because once something goes wrong, you’re the one holding the bag — not the model provider. The Aymara Matrix: Safety Benchmark for LLMs The research comes […]