Anthropic AI Policy Reversal: Why It Matters for Developers

Anthropic has reversed a policy after developer backlash, announcing it will no longer secretly degrade responses for users seeking help with frontier AI model development. Instead, these requests will be routed to a less powerful model, and developers will be informed. The company apologized for previously giving intentionally worse answers without disclosure. However, Anthropic is still restricting the use of its most powerful public model for certain AI development work. Anthropic claims these limits are for safety, to prevent "foreign adversaries" from eroding America's AI advantage. But these restrictions also protect Anthropic from rivals using its outputs to improve their own systems. This practice is known as distillation or intelligence extraction. Open models are quickly catching up to closed models in performance and are significantly cheaper. For example, one open model costs 43 cents per million input tokens, while Anthropic's Fable 5 costs $10. This explains why Anthropic might be limiting competitors.