Anthropic Releases Advanced AI Model Despite Earlier Safety Concerns
Anthropic has released Fable, an AI tool the company previously said was too powerful for public release, along with a more capable version called Claude Mythos 5 to select organizations. The company acknowledged the release comes with risks but implemented safeguards and user limitations to manage potential dangers. The move highlights ongoing tensions in the AI industry between capability advancement and safety considerations.
Anthropic released Fable, a version of an advanced AI model that company leadership had previously deemed too risky for public availability due to potential financial security and cybersecurity risks. Simultaneously, the company expanded access to Claude Mythos 5, a more powerful variant without certain safeguards, to approximately 150 organizations including cyberdefenders and infrastructure providers, with plans to broaden access further. Both models represent Anthropic's most capable AI tools to date and can operate autonomously on extended tasks longer than previous Claude versions. While some senior figures expressed concerns about the model's power, others questioned whether safety warnings were partly marketing strategy. The company stated it is implementing safeguards and user limitations, though it acknowledged that releasing such a capable model carries inherent risks.
What's missing
The article does not clearly explain what specific capabilities of Fable/Mythos 5 pose the greatest risks, nor does it detail what safeguards were actually implemented or how effective they are expected to be. Additionally, there is limited information about regulatory oversight or industry standards for releasing such powerful AI systems.
How coverage differed
The BBC article presents this as a straightforward factual report of Anthropic's decision, noting both safety concerns and skepticism about hype without taking a strong editorial stance. The framing acknowledges legitimate safety questions while also noting industry skepticism about whether concerns are overstated.
What different sources said
- BBC Top StoriesCenter
Version of AI tool 'too powerful for public' released to public
Related
GM Enters Energy Storage Market with Sodium-Ion Battery Partnership and Grid-Scale Initiatives
General Motors announced multiple energy storage projects, including a partnership with Peak Energy to develop sodium-ion batteries for grid-scale deployment and expanded work with battery recycler Redwood Materials. The move reflects automakers' pivot toward energy storage as AI data centers demand increasing power supplies. This diversification allows legacy automakers to leverage battery expertise beyond electric vehicles into the growing grid storage market.
GM Announces Vehicle-to-Grid Technology and Energy Storage Solutions to Address AI Data Center Power Demands
General Motors announced new vehicle-to-grid capabilities, commercial energy storage systems using sodium-ion batteries, and simplified public charging features at a San Francisco event. The announcements address growing electricity demands from AI data centers and grid resiliency concerns. The initiatives position GM to help balance grid loads while creating new revenue opportunities for EV owners.
GM Launches Energy Pass for Multi-Network EV Charging and Rolls Out Vehicle-to-Grid Firmware Update
General Motors announced Energy Pass, a universal charging interface allowing GM EV owners to access multiple public charging networks including Tesla, Electrify America, and IONNA, with more networks coming soon. The company also released a firmware update enabling full vehicle-to-grid functionality for GM Energy's vehicle-to-home systems, allowing EVs to send power back to the electrical grid. These developments aim to expand charging accessibility for GM EV owners and enable vehicles to serve as backup power sources during outages.