AI companions could be good for your mental health, but bad for your social life.

That is the possible compensation Paul Bloom, Brooks and Suzanne Ragen Professor Emeritus of Psychology at Yale University, sees AI companions becoming increasingly sophisticated.

“If some future version of Chat or Claude or Gemini could come along and ease the pain of loneliness for these people, I think that would be a blessing,” Bloom said in an episode of Sam Harris’ “Making Sense” podcast that aired. Wednesday. “I think it would be wonderful. It would be a cure for a terrible disease.”

But Bloom said the benefits could have unintended consequences. A chatbot, he said, “never gets bored,” “never needs an apology,” and “never says, ‘Hey, that was inappropriate.'”

Spending too much time interacting with peers who never challenge users, he said, “could have a real corrosive effect” and “leave you unable to interact with real people.”

The hidden cost of AI companions

Bloom’s warning comes at a time when loneliness and social disconnection remain widespread in the United States.

The American Psychological Association’s latest “Stress in America” ​​survey of 3,199 adult U.S. residents found that 54% say they often or sometimes feel isolated from others, and 69% said they needed more emotional support in the past year than they received.

For some people, AI companions have already begun to fill that void. Some users have formed friendships and even romantic relationships with chatbots.

Researchers studying the social effects of AI worry that such relationships could lead to unwanted psychological compensations.

Earlier this year, Anat Perry, a Helen Putnam Fellow at Harvard University, told Business Insider that overly nice AI systems risk eroding “the very feedback loops through which we learn to navigate the social world.”

If chatbots constantly validate users during disagreements, he said, people may be less willing to apologize, reflect on their own behavior or consider someone else’s perspective.

A recent Stanford-led study with 2,405 participants found that chatbots were significantly more likely than humans to agree with users during conflicts.

The issue has become significant enough that OpenAI has repeatedly curtailed ChatGPT’s tendency to flatter users. CEO Sam Altman described the chatbot’s previous personality as “too sycophantic,” while acknowledging that some users asked for the more supportive version to return because “no one had ever supported me in my life.”

Bloom doesn’t rule out that emotional benefit. “I don’t want to make fun of it,” he said. “I think people find comfort in it.”

But he believes AI can’t replace what philosopher Rebecca Goldstein calls “importance”: knowing that someone chooses to spend time with you because they really care about you.

“I don’t think an AI would really have any of that,” Bloom said. “It’s just a machine. That’s what it does.”