OpenAI CEO Sam Altman has criticized rival Anthropic for allegedly using “fear-based marketing” to promote its Claude Mythos model, which is noted for its capability to identify software vulnerabilities and execute advanced cyber operations. Altman believes this approach is meant to limit access to AI technology, keeping it in the hands of a select few, despite there being valid safety concerns regarding AI. Claude Mythos has recently been under scrutiny as the National Security Agency begins testing it on classified networks, reflecting broader industry debates over controlled versus open access to powerful AI systems, with Anthropic choosing a more restricted distribution strategy.

OpenAI: OpenAI is an AI research organization that builds advanced generative models including the GPT series and operates ChatGPT. In response to Anthropic’s Claude Mythos rollout, OpenAI CEO Sam Altman criticized the approach as fear-based marketing designed to limit AI access to a trusted few. OpenAI favors broader distribution of powerful AI to drive innovation while acknowledging legitimate safety concerns.
Anthropic: Anthropic is an AI safety and research company that develops the Claude family of large language models with a focus on reliability and controlled deployment. The company recently introduced Claude Mythos, a model capable of identifying software vulnerabilities and simulating cyber operations, which it distributes exclusively through Project Glasswing to select partners like Amazon, Apple, and Microsoft. Anthropic promotes the model as a defensive tool while restricting access to address potential offensive risks.
Sam Altman: Sam Altman serves as CEO of OpenAI, leading efforts in scaling AI infrastructure and model development. On the Core Memory podcast, he accused Anthropic of using fear-based marketing to promote Claude Mythos by exaggerating dangers to justify restricted access. Altman emphasized OpenAI’s commitment to responsibly releasing advanced capabilities despite industry debates on control.

Government Testing: The National Security Agency is testing Claude Mythos on classified networks amid concerns over warfare and surveillance applications.
Mythos Capabilities: Claude Mythos demonstrates prowess in vulnerability detection, such as in Firefox, and multi-stage cyber simulations.
AI Deployment Divide: The AI sector debates controlled versus open access to powerful models, with Anthropic opting for restrictions and OpenAI advocating wider availability.