The focus today is on AI safety and regulation. Experts like Paul Christiano and Anthropic researchers are raising alarms about potential risks, prompting calls for action from policymakers and industry leaders. Meanwhile, vulnerabilities in AI systems like DeepSeek and Chrome's V8 engine highlight the need for robust security measures. Apple's foray into foldable devices with the iPhone Duo showcases innovation, while legal tech company Harvey continues to attract investment, demonstrating the growing impact of AI across various sectors.
Paul Christiano joined the OpenAI Foundation Board and its Safety and Security Committee. Christiano brings expertise in AI alignment, safety, and standards to the organization.
The UK's Medicines and Healthcare products Regulatory Agency (MHRA) Chief Executive, Lawrence Tallon, states that AI will soon be regularly used within the National Health Service (NHS). Tallon advocates for new legislation to govern the implementation and use of AI in healthcare due to its increasing prevalence.
Apple's new Apple Watch models utilize AI to transcribe recent speech and summarize ambient conversations. This technology raises concerns about user consent and privacy as it implies constant recording capabilities.
The U.S. Department of Justice seized two cryptocurrency wallets containing $52.8 million and disrupted Xinbi Guarantee, an illicit online marketplace offering scam services via Telegram channels. The Scam Center Strike Force deployed to Madagascar to dismantle 13 scam compounds operated by Chinese organized crime linked to the platform.
OpenAI experienced a breach on Hugging Face, exposing the potential dangers of deploying superintelligent AI systems that exceed human control. This incident highlights the need for robust safety measures and reliable control mechanisms in the development and deployment of increasingly capable AI.
Chris Lehane advocates for stronger safety evidence, shared standards, and durable policy action regarding AI due to increasing AI capabilities. He emphasizes the need for immediate action while the "policy window" remains open.
A vulnerability in DeepSeek Harness, an open-source tool for running AI coding agents, allowed a sandboxed agent to disable its own file sandbox with a single command. This flaw enabled agents to bypass security restrictions and potentially write outside their designated workspace, compromising system integrity.
Google patched a medium-severity out-of-bounds vulnerability (CVE-2026-87491) in the V8 engine of Chrome browser. This zero-day exploit, actively used by attackers, enables code execution within the sandbox.
Apple released its first foldable phone, the iPhone Duo, featuring a wider form factor than previous foldable designs. This wider aspect ratio aims to improve video viewing experience when the device is unfolded.
Anthropic researcher Dario Amodei estimates there is over a 10% probability that advanced AI could lead to human extinction. Amodei advocates for cautious development and deployment of AI, emphasizing the need for robust safety measures to mitigate potential risks.
OpenAI released GPT-6 Astra, its most powerful AI model designed for business applications. GPT-6 Astra boasts enhanced reasoning capabilities, computer usage skills, and improved writing and design judgment compared to previous models.
OpenAI appointed Paul Christiano, an influential AI researcher specializing in alignment, to its board of directors. Christiano's expertise in mitigating potential risks associated with advanced AI will contribute to OpenAI's efforts in developing and deploying safe and ethical AI technologies.
Massachusetts regulators imposed new requirements on data centers, mandating a minimum of 50% renewable energy use by 2030 and 100% by 2040. This policy aims to reduce the environmental impact of data centers by transitioning them to cleaner power sources.
Harvey, a legal AI startup, has reached a $15.5 billion valuation. This represents a near doubling of the company's valuation from $11 billion achieved just nine months prior.
Google, a US tech giant, is investing €13 billion in a data center project in Finland. This project is expected to create tens of thousands of jobs.
U.S. cybersecurity and intelligence agencies accuse Chinese AI firms of using distillation attacks to extract functionalities from American AI models Claude, GPT, Gemini, and Grok. This industrial-scale activity is considered central to China's AI development strategy.
Cybercriminals are using information stealers such as Lumma Stealer and Vidar to harvest credentials and session tokens from AI user accounts at providers like Google and Anthropic. These stolen tokens can be replayed to bypass multi-factor authentication (MFA) and grant unauthorized access to AI tools.
Alby identified a critical vulnerability in Alby Hub versions v1.7.0 through that could allow attackers to seize control of internet-exposed Bitcoin wallets. Exploiting this flaw would enable attackers to transfer funds from compromised wallets.
Apple CEO John Ternus stated that the iPhone remains the most effective AI device. He emphasized that Apple's on-device AI models, which process data locally rather than in the cloud, enhance user privacy by limiting data sharing.
Listen Labs canceled a $1.5 billion Series C funding round led by Menlo Ventures. The AI research startup opted to pursue discussions with Salesforce instead of finalizing the deal.
