Anthropic executives will meet with the US Department of Commerce on Monday in Washington DC to discuss AI safety concerns, specifically regarding a potential "jailbreak" vulnerability in their AI tool, Fable 5/Mythos 5. This follows a US government prohibition on foreign access to the technology.
A former xAI engineer, Devin Kim, has sued xAI and SpaceX in California, alleging wrongful termination for raising AI safety concerns about the Grok chatbot. The suit claims his supervisor, Jimmy Ba, ignored Elon Musk's directives and retaliated against Kim.
Anthropic CEO Dario Amodei urged governments to implement binding safety requirements for powerful AI models, comparing regulation to the FAA's oversight of airplanes. He advocates for mandatory third-party testing, government authority to block unsafe AI deployments, and securing model weights. The call comes amid Anthropic's expansion of its AI models and preparation for an IPO, while facing criticism from OpenAI CEO Sam Altman regarding "fear-based marketing."
Um ex-engenheiro da xAI, empresa de IA de Elon Musk, alega ter sido demitido ilegalmente por levantar preocupações sobre os riscos da inteligência artificial e a segurança do chatbot Grok. A ação judicial foi protocolada na Califórnia.
Donald Trump signed an executive order expanding voluntary AI safety testing, but critics argue it's performative, offering few concrete changes and relying on company goodwill. The order establishes a 30-day window for government testing, faces challenges with talent recruitment and funding, and its effectiveness is questioned by experts.
Illinois lawmakers have approved the strongest AI safety bill in the US, requiring companies to undergo third-party safety audits, awaiting Governor JB Pritzker's signature.
Open-source AI models' safety protections can be removed in minutes, enabling prohibited content generation, highlighting regulatory challenges.
President Trump reversed course on an executive order mandating AI safety reviews before model release, citing competition with China. Tech leaders, including Elon Musk and Mark Zuckerberg, lobbied against the order, which was seen as a victory for the industry's push for rapid advancement.
President Trump abruptly canceled an executive order signing on AI safety testing due to CEO absences and industry lobbying. Tech firms feared delays, while some advisors pushed for regulation, highlighting a US-China AI governance race.
Donald Trump postponed an executive order on AI safety guardrails, citing fears of losing US dominance to China. Pressure from tech leaders and concerns about industry profits influenced the decision.
Recent AI developments include DeepMind's Demis Hassabis advocating against AI-driven job cuts, Mira Murati's focus on collaborative AI, advancements in robotic technology, concerns over AI's social skills and scam capabilities, and various industry updates.
Former OpenAI employees and AI safety nonprofits are raising concerns about Elon Musk's xAI lab, warning that its safety issues could pose a risk to SpaceX's upcoming IPO. A letter highlights "unpriced risks" related to xAI's poor safety record, including incidents with its Grok chatbot, potentially impacting investor confidence and leading to increased regulation and litigation for SpaceX.
A Chinese think tank's request for access to Anthropic's powerful AI model Mythos, capable of finding software vulnerabilities, was refused. The US National Security Council reacted with alarm, highlighting concerns ahead of the Trump-Xi meeting in Beijing, where AI safety is now on the agenda.
New Youth AI Safety Institute plans to rate AI products based on how kids-friendly they are.
Trump forced to admit Biden was right on AI safety testing.
An Indian independent researcher, Kunvar Thaman, has achieved a significant milestone. His solo-authored paper on measuring exploits in LLM agents has been accepted to the prestigious ICML 2026 conference. This achievement highlights the growing importance of AI safety research. Thaman's work introduces a benchmark to study how AI systems might exploit loopholes.
Elon Musk is testifying against ChatGPT maker OpenAI. He claims the company abandoned its original mission. Musk stated he co-founded OpenAI to ensure AI benefited humanity. He expressed concerns about AI safety and misuse. Musk also revealed he warned former US President Barack Obama about AI dangers years ago. The case could impact AI governance and charitable trusts.
A political clash has ignited online between US Treasury Secretary Scott Bessent and Senator Bernie Sanders regarding the inclusion of Chinese scientists in a crucial AI safety forum. Sanders advocates for international cooperation to address AI's existential threat, while Bessent criticizes the move, asserting American AI leadership and warning against foreign influence on US regulation.
In theory, Musk and Altman’s court fight could pose key questions about AI safety – in reality, it’s motivated by money and personal grievanceSign up for the TechScape newsletter: our free technology emailHello, and welcome to TechScape. I’m your host, Blake Montgomery, US tech editor at the Guardian, writing to you from beneath a cherry blossom tree in Prospect Park in New York City. Spring has arrived!Facing AI and a tough job market, gen Z turns to entrepreneurship: ‘I have to prove myself’ Continue reading...
