'AI not your friend': LinkedIn co-founder Reid Hoffman says you will be harmed if you pretend so

LinkedIn cofounder and prominent AI investor Reid Hoffman is raising concerns about a growing trend in the tech industry: positioning AI systems as emotional companions or “friends.”

In a recent episode of the Possible podcast, as reported by Business Insider, Hoffman warned that no AI system today is truly capable of being a friend, and pretending otherwise may actually harm users. “I don’t think any AI tool today is capable of being a friend,” he said. “And I think if it’s pretending to be a friend, you’re actually harming the person in so doing.”

His remarks come amid Meta’s aggressive push to roll out AI companions across Facebook, Instagram, WhatsApp, and Ray-Ban smart glasses. CEO Mark Zuckerberg recently told podcaster Dwarkesh Patel that AI chatbots could help address America’s “loneliness epidemic,” citing research suggesting many Americans have fewer than three close friends, despite the capacity for more.

But Hoffman drew a firm line between companionship and genuine friendship, warning that blurring those definitions could dilute the meaning of human connection. “Friendship is a two-directional relationship,” he said, explaining that a real friend both offers and expects emotional support and accountability. “It’s not only, ‘Are you there for me?’, but I am here for you.”

Hoffman stressed that while AI systems might simulate empathy, they lack the human reciprocity that defines true relationships. “Calling AI your friend is dangerous because it changes how we understand real human relationships,” he said.

He praised the design of Inflection AI’s “Pi” assistant, which clearly states that it is a companion, not a friend, and encourages users to maintain real-world human connections. “Helping you go out into your world of friends is, I think, an extremely important thing for a companion to do,” Hoffman said.

While acknowledging that AI can be helpful, especially for those with limited social interaction, Hoffman cautioned against letting it replace genuine interpersonal relationships, particularly for children.

Hoffman argued for clearer regulation and accountability in the industry. “We as a market should demand it, we as an industry, all MPAs, should standardise around it,” he said. “And if there’s confusion around this, I think we as government should say, ‘Hey, look, if you’re not stepping up to this, we should.’”

For Hoffman, the stakes are existential. “I think that’s a degradation of the quality of elevation of human life,” he said. “And that should not be what it’s doing.”

You can watch the episode here.