AI chatbots are becoming part of everyday life - not just for keeping track of All Whites v Belgium or practical tasks, but for companionship, advice, and emotional support.
There’s a lot of excitement about this, but also some real concerns. So today I thought I’d talk about what AI does well, where it falls short, and what that means in practice.
What AI is good at…
It’s always there: One of the biggest advantages is availability. AI is 24/7 and immediate. So if someone is struggling at 2am, they don’t have to wait, they can talk to something straight away. For people who feel embarrassed or unsure about asking for help, that low barrier can be really important.
It can also make people feel heard (at least in the short term). Even though AI isn’t truly empathic, people often feel understood and supported when they talk to it. That can help reduce negative emotions, give people a sense of reassurance, and provide some sense of validation i.e., “this thing gets me”. Part of that is because people see AI as safe and non-judgmental which are key factors in helping people open up.
Another strength is helping people prepare for real conversations. People can use AI to sort out what they’re feeling and practice what they want to say. So it can act like a kind of thinking tool or rehearsal space. My daughter used it for this reason last year and found it really helpful to think about whether she should take the next step and contact a real person.
However, there are some more tricky things about using AI for advice and emotional support.
Firstly, while it can help people feel heard and understood in the moment, new research from the UK suggests that in the longer term using AI for advice and emotional support leads to people feeling more lonely compared to not using it. It seems that, over time, people can come to prefer using AI instead of talking with other humans, but that the type of interaction you get with AI is not as satisfying as talking to a real person. But we can end up getting trapped into the loop of using AI and slowly become out-of-practice discussing these issues with real people.
It can also give a false sense of progress, sometimes called “false completion”. AI can make people feel like they’ve “worked something out” - but not really. We get a sense of false completion because we’ve talked something through and consequently feel a bit better but nothing actually changes, which can delay people seeking real help for themselves. Relatedly, we tend to treat AI chatbots like humans, when obviously they’re not, which can make our interactions feel more meaningful than they actually are.
Overall, AI is best seen as a support alongside, but not instead of, human connections - particularly when you might need some support at unusual times or want to get your own thinking straight. But beware of the trap of coming to rely on it instead of real human connections, as AI can’t replicate what we get from human-to-human relationships.
LISTEN ABOVE
See omnystudio.com/listener for privacy information.
Jun 27 20266 mins
Jun 27 20261 hr and 57 mins