Neither company is acting responsibly
Misleading
The evidence is mixed. Anthropic loosened its core safety principle in response to competition, adopting a nonbinding safety framework instead of self-imposed guardrails, citing shortcomings in its Responsible Scaling Policy could hinder its ability to compete. However, OpenAI and Anthropic conducted a joint safety evaluation, testing each other's models for misalignment, instruction following, hallucinations, and jailbreaking, and leading AI companies are increasingly aligning on efforts to strengthen AI safety through tighter controls and coordinated standards. This shows mixed responsibility, not uniform irresponsibility.
This claim may mislead readers without added context.
OpenAI and Anthropic are racing straight to self-improving superintelligence
Misleading
Both companies are using their own models to accelerate research, with Anthropic reporting that the majority of its code is now written by Claude, and OpenAI reporting that GPT-5.5 developed a parallelization method that boosted token generation speeds. However, OpenAI, Anthropic, and Google DeepMind all use AI to assist their research workflows, but their businesses are built around selling models and API access, not organized around recursive self-improvement as a core commercial thesis. The statement overstates their singular focus.
This claim may mislead readers without added context.
They are gambling with our lives
Can’t verify
This is a subjective value judgment about risk assessment and acceptable tradeoffs. While people building AI fear progress will soon accelerate violently, whether this constitutes 'gambling with lives' depends on one's risk tolerance and beliefs about AI safety—not verifiable facts.
Add to Chrome