Research Finds Limitations of ChatGPT in Psychotherapy


Meta's "Cannes" project hires outsiders posing as minors to stress-test rival chatbots like ChatGPT, Gemini, and Character.AI, running at least until April. The aim is to probe AI safety limits, but the deception sparks controversy.....
AI star Anthropic is building a large data center in Hawesville, Kentucky, signing a 20-year lease with TeraWulf. Located about an hour's drive from Louisville, this move marks the Claude-powered unicorn's accelerated push into foundational computing, creating a supercomputing hub of unprecedented scale.....
ChatGPT had a security vulnerability where attackers could use prompt injection and path traversal to bypass file access restrictions and obtain unauthorized data. The flaw stemmed from mishandling of uploaded files: while the system didn't offer raw file downloads, the exploit allowed stealing extra information. OpenAI has fixed it.....
Japan announced new guidelines for its "AI Robot Strategy", planning to deploy 10 million AI robots in 18 industries including manufacturing, healthcare, and agriculture by 2040, to systematically address the labor crisis caused by low birthrate and aging population.....
Alibaba issued an internal ban to discontinue all Claude products from Anthropic by July 10, including models like Sonnet, Opus, Fable, and agents like Claude Code. Previously, Alibaba had allowed external AI use with free quotas and reimbursements to accelerate intelligence, hinting at a strategic shift.....