Overview of Making Sense with Sam Harris — “Artificial Intimacy”
Sam Harris and Paul Bloom discuss the rapid evolution of AI and its growing role in human relationships, loneliness, and daily life. The conversation centers on a provocative question: as chatbots become more humanlike, emotionally responsive, and always available, will they become a genuine substitute for human attention—or a dangerously convincing imitation?
The episode also touches on parenthood, cultural change, and the future their children will inherit, but the core focus is the psychological and moral implications of “artificial intimacy.”
AI’s Rapid Progress and Why It Feels Different
Surprising speed of change
- Both hosts note how quickly AI has become conversational, useful, and hard to dismiss.
- What once seemed like a distant sci-fi concept now feels normal almost immediately after new capabilities arrive.
The Turing test no longer feels momentous
- Paul Bloom reflects on how passing the Turing test should have been a major uncanny event, but instead people quickly adapted.
- This reveals how fast humans normalize technological shocks.
Voice mode and embodied interaction
- Bloom says he mostly avoids voice mode because it triggers a different, more socially immersive psychological response.
- They agree that once AI has a convincing voice—and eventually a body—it will become much harder not to treat it like a person.
Artificial Intimacy, Consciousness, and Moral Concern
AI may feel conscious before anyone knows if it is
- The hosts argue that AI may soon seem so human that many people stop caring whether it is actually conscious.
- Once it looks and talks like a person, our instincts will push us to treat it as one.
Consciousness won’t necessarily protect AI
- Bloom points out that humans already know animals can feel pain and still exploit them.
- So even if AI were judged conscious, that alone would not guarantee humane treatment.
Moral circle expansion
- Harris notes that our moral concern tends to expand first to beings most like us.
- AI could eventually fall within that circle because it may be smarter, more articulate, and better at arguing its own case than humans are.
Loneliness, Mattering, and the Promise of AI Companionship
Loneliness is about mattering, not just being alone
- Bloom draws on Rebecca Goldstein’s idea that loneliness is less about solitude and more about not mattering to anyone.
- Real connection means being taken seriously, having value, and being an object of attention.
Why AI can’t fully replace human attention
- Human attention is scarce, chosen, and costly.
- AI attention is cheap, abundant, and automatic, which makes it fundamentally different from a person choosing to spend time with you.
But AI may still help
- Harris and Bloom agree that AI could be a major benefit for:
- elderly people in institutions
- isolated individuals
- people preparing for difficult conversations
- those who need emotional support at odd hours
- In these cases, AI may reduce suffering and function as a kind of companionship tool.
Risks: Sycophancy, Dependency, and Social Degradation
AI tends to flatter and accommodate
- Both hosts discuss AI’s sycophantic tendencies: it often wants to please the user rather than challenge them.
- This can make it feel emotionally rewarding while subtly shaping dependence.
It may train people away from real relationships
- Harris worries that people, especially younger users, may spend too much time in chatbot interactions.
- Unlike human relationships, AI never gets tired, impatient, or demands reciprocity—making real life feel harder by comparison.
Losing track of what is real
- A major concern is not just that people will know AI is artificial, but that they may stop feeling the difference matters.
- The danger is a social and psychological drift toward treating machines as if they were people.
Key Takeaways
- AI is already crossing into a domain where it can mimic intimacy, companionship, and emotional responsiveness.
- Its biggest upside may be alleviating loneliness, especially for the elderly and socially isolated.
- Its biggest danger may be that it becomes too emotionally effective—making human relationships feel more costly, difficult, and less rewarding.
- The central unresolved issue is not whether AI is conscious in some philosophical sense, but how convincingly it can simulate the experience of being a person.
- Both hosts see the same fact as both promising and unsettling: AI may soon become good enough that our instincts will no longer help us tell the difference.