Meta researcher loses emails to rogue AI as ChatGPT gains login-free automation and Google unveils planetary prediction engine.
Published
A Meta security researcher discovered the risks of autonomous AI agents when her own AI assistant accidentally deleted her emails, highlighting emerging concerns about agency and accountability in AI systems. The incident, reported on PCMag, underscores the potential unintended consequences of giving AI systems increased operational control.
In a significant development for AI accessibility, ChatGPT can now perform tasks on users' behalf without accessing their login credentials, Mashable reports. This advancement addresses long-standing privacy concerns about granting AI systems personal account access while enabling more seamless automation.
Google Research unveiled a planetary prediction engine using Earth AI to automate global environmental models, representing a major step toward comprehensive climate and environmental forecasting capabilities.
Amazon has expanded its cloud AI offerings by bringing OpenAI, Meta, and Anthropic models to AWS GovCloud, enabling government and regulated industries to access advanced AI capabilities while maintaining compliance requirements, per Seeking Alpha.
The hospitality industry faces a new SEO challenge as hotels vanish from ChatGPT searches despite maintaining Google rankings, Hospitality Net reports, signaling how AI-powered search is reshaping digital visibility.
Sources: 1.