Anthropic Details How It Measures Claude’s Wokeness

Why It Matters

The initiative signals a broader industry push for politically neutral AI, which could shape government procurement standards and set new benchmarks for bias mitigation in commercial large‑language models.

Summary

Anthropic unveiled new safeguards to make its Claude chatbot politically even‑handed, embedding system prompts that forbid unsolicited political opinions and using reinforcement learning to reward neutrality. The company also released an open‑source tool that scores Claude’s responses, reporting 95% even‑handedness for Claude Sonnet 4.5 and 94% for Claude Opus 4.1, ahead of Meta’s Llama 4 (66%) and OpenAI’s GPT‑5 (89%). These steps follow heightened political scrutiny, including a Trump executive order demanding unbiased AI for government use and similar bias‑reduction pledges from OpenAI. Anthropic acknowledges the approach isn’t foolproof but expects it to substantially improve balanced discourse.