Anthropic's latest advertising campaign, which featured voices from individuals interviewed during a U.S. roadshow, has drawn criticism for appearing tone-deaf. The campaign was designed to demonstrate that the artificial intelligence company is actively listening to public thoughts, feelings, worries, and excitements about AI, according to Daniela Amodei, Anthropic's president. Amodei stated that this initiative was a culmination of efforts to acknowledge the validity of public concerns and to engage with them directly as a leader in AI model training and development.
The backlash to the campaign underscores a broader challenge: the dismissal of AI industry warnings about potential risks as mere marketing tactics. Anthropic previously decided not to release its Mythos Preview, citing its cybersecurity capabilities as too dangerous. While governments took this warning seriously, much of the public perceived the announcement as a sophisticated marketing stunt. This perception persists despite evidence, such as an incident where a less powerful version of Anthropic's Claude AI was used to help steal 150 gigabytes of data from the Mexican government, indicating Mythos Preview's significantly greater power by independent measures.
The company has conducted extensive outreach, surveying over 120,000 people across 159 countries as part of its broader engagement efforts. The recent ad campaign was intended to reflect these engagements, aiming for a more candid approach than typical AI marketing, though it ultimately struck many observers as disconnected from public sentiment.