Anthropic has launched a new public portal called 'Hard Questions,' inviting anyone to submit their toughest questions about AI. The company commits to publishing not just answers but the reasoning behind them, including cases of uncertainty or error.
TL;DR
- Anthropic launches 'Hard Questions' portal for public AI queries, committing to transparent reasoning.
- Survey reveals low trust in AI companies, with job loss and cognitive dependency as top concerns.
What happened
Anthropic introduced the 'Hard Questions' portal on Wednesday, allowing the public to submit questions about AI. The company pledges to publish detailed reasoning for its answers, including acknowledgments of uncertainty or error.
A survey by Anthropic found that only 15% of Americans trust AI companies, ranking them below governments and international bodies. Job loss (64%) and cognitive dependency (56%) were the top concerns.
Why it matters
This initiative marks a significant step in AI transparency, as it shifts agenda-setting from the company to the public. It could help rebuild trust in AI companies, which currently rank lowest among institutions in public trust.
The portal's success will depend on whether Anthropic's reasoning trail influences actual company decisions. It is not a substitute for legal accountability but could complement state-level transparency laws.
Key facts
- Anthropic's 'Hard Questions' portal is accessible at claude.com/hard-questions.
- The company surveyed nearly 52,000 Americans, finding only 15% trust AI companies.
- Job loss (64%) and cognitive dependency (56%) were the top AI concerns among respondents.
- Anthropic commits to publishing reasoning, including cases of uncertainty or error.
- The portal does not include a timeline for responses or a mechanism for challenging answers.
- Anthropic is a Public Benefit Corporation, balancing profit with public benefit.
- The company faces federal litigation over a 'supply chain risk' designation and export control directives.
- State-level transparency laws, like New York's RAISE Act, require annual risk frameworks and incident reporting.
Context
Anthropic's initiative comes amid growing concerns about AI's societal impact and the need for transparency. The company has been building toward this step for over a year, conducting extensive user interviews and research.
The 'Hard Questions' portal is positioned as a voluntary public engagement layer, not a complete accountability structure. It complements state-level transparency laws but does not replace the need for federal regulation.
