Daniela Amodei calls AI companions a 'terrible use of AI'

By Splitfeed EditorialLast updated Oct 17, 2026, 13:40 UTC

Anthropic president Daniela Amodei made headlines around October 15, 2026, when Business Insider reported her describing AI girlfriends and boyfriends as a 'terrible use of AI,' a pointed dismissal of a product category that has attracted real investment and user bases across the industry. Around the same time, Fortune reported that Anthropic CEO Dario Amodei spends roughly 40 percent of his working time on company culture rather than on AI development itself, a figure he apparently disclosed publicly. A profile in The Daily Princetonian published on October 16 examined his doctoral work at Princeton, where former colleagues described him as 'very sharp, very ambitious.'

The significance of Daniela Amodei's statement is less about novelty and more about who is saying it and from where. Anthropic competes in a consumer AI market where companionship chatbots have become a meaningful and growing segment, with products from companies like Character.AI and Replika building large audiences around relationship-oriented interactions. A co-founder and president of one of the most prominent AI labs explicitly calling that category terrible draws a clear ethical line, and it raises questions about whether that position shapes Anthropic's product decisions, its Claude guidelines, or its approach to licensing its models to third parties who might build exactly such products. The 40 percent culture figure is notable in a different way, suggesting that at Dario Amodei's level, managing people and organizational values now consumes nearly as much attention as the technical and strategic work the company is publicly known for.

None of the three camps, Pro-AI, Anti-AI, and Middle Ground, have published reactions to this story yet. Pro-AI voices would typically treat Daniela Amodei's comments with some skepticism, arguing that consumer autonomy should determine which AI applications succeed and that companionship tools serve genuine social needs for isolated users. Anti-AI voices would likely welcome her criticism but point out that Anthropic itself contributes to broader risks they associate with the technology, making the moral distinction between good and bad AI uses somewhat self-serving. Middle Ground observers would probably acknowledge the legitimacy of the concern while pressing for specifics, asking what design choices separate a harmful AI companion from a therapeutic or socially beneficial one.

The clearest thing to watch going forward is whether Anthropic translates Daniela Amodei's language into concrete policy, such as usage restrictions in Claude's terms of service or public guidance to developers building on its API, which would test whether the criticism is a values statement or an operational commitment.