Overview
- This week Anthropic released a 90‑second 'Hard Questions' ad for its Claude chatbot that pairs recorded questions from people with stark images such as protests, the unhoused, burning buildings and military tombs.
- The ad has produced mixed reactions online, drawing praise for confronting risks and sharp criticism for tone and choice of imagery, with OpenAI CEO Sam Altman mocking the clip on X.
- Anthropic says the spot grew from a U.S. roadshow and a broader survey effort that gathered responses from more than 120,000 people across 159 countries.
- Critics frame the ad as evidence of a deeper credibility problem because Anthropic has previously withheld a powerful model, Mythos Preview, for security reasons and reporting has linked a weaker Claude instance to a reported 150 GB data theft involving the Mexican government.
- The campaign intensifies scrutiny of Anthropic’s safety messaging as it prepares for larger commercial moves and faces regulatory pressure, and observers will watch whether the company’s warnings win trust or are seen as brand positioning.