S05E98 The Empathy Edge: What AI Chatbots Teach Us About Listening

In a recent Wall Street Journal article , he talks about how those large language models yes , such as chat GPT that you've been hearing so much about are actually outdoing humans at making people feel heard and understood Intriguing right . Well , let's set an example scene .

Picture this You've taken a break from work for a few years to start a family , and now you're ready or maybe not so ready to jump back into your career . You're anxious , you're feeling unmotivated , and you decide to share your worries with your friends . And your friends respond , and friend one says I'm sorry to hear that you're struggling .

I can understand how anxiety and insecurity can make it hard to take that next step . You have valuable skills , though , and experience that can benefit any employer , and you deserve to feel fulfilled in your career . Friend two says oh , I've struggled with the same problem and the best way to tackle it is to jump right in and give it your best .

Now ask yourself which response felt more compassionate and more attentive . Of course , chances are you picked friend one and , of course not surprisingly because of this topic it's actually from a chatbot .

So how big is this empathy gap between humans and AI ? Well , first of all , it's not anecdotal . There are multiple studies where people have rated chatbot responses as more caring , more empathetic and more helpful than those from humans .

For example , a study in JAMA Internal Medicine showed that patients with medical concerns prefer chatbot replies over a real doctor's response a whopping 80% of the time . 80% of the time . And while , as a physician myself , my collective physician ego is probably hurt , since I've worked a lot with other physicians , I'm actually surprised that it's not more often .

In another study from communication psychology , researchers found crisis chatbot responses were seen as more compassionate than those from trained hotline crisis responders , and so this raises a big question how on earth are chatbots beating humans on empathy ? They're not even human and , of course , they can't even feel emotion .

So why might chatbots seem to be better listeners ? Well , the first theory that researchers had was that chatbots have virtually unlimited time , so they could devote themselves fully to each person's issue . Meanwhile , busy physicians or crisis line volunteers are stretched thin . But the research study showed that time alone isn't the real reason for AI advantage .

In a 2024 paper from Harvard Business School , participants who were incentivized to provide thoughtful , empathetic responses still couldn't match ChatGPT's level of perceived compassion , even though these subjects spent more time trying .

So if it's not time or incentives , what is it ? According to the research , chatbots simply avoid the common mistakes we humans make . When we try to help , we focus on ourselves . We shift the spotlight by sharing our similar experience , even if we're just trying to relate . We jump right to problem solving .

We rattle off solutions without first validating how the other person feels . Chatbots , by design , don't do this . They have no personal stories to tell , no eagerness to fix your problem , no ego to protect . Instead , they take the time to paraphrase what you've said , showing that they're paying attention . They validate your feelings .

They ask follow-up questions to dig deeper into your experience . Basically , they make it all about you . What are the pitfalls of our good intentions ? Let's be honest when someone we care about is hurting , our first instinct is often to either

cheer them up or hand them advice . I know many people close to me who do exactly that . Of course we mean well , but it can come off as dismissive . It might also minimize the other person's struggle or shift the conversation onto their own experiences .

Chatbots , on the other hand , don't need to prove themselves or share a personal anecdote , or even skip to a bunch of solutions . They effortlessly keep the focus where it belongs on the person who needs support . Zaki's article describes one technique that makes a difference , called looping for understanding .

You basically repeat what the person said in your own words , then check to see if you've captured it correctly . Do I have that right ? It shows you're truly listening .

This is a technique that we actually learned in our first patient communications course way back in medical school , and I think Dale Carnegie also uses a similar technique when teaching people how to speak to others . Interestingly , chatbots do this naturally , and humans need to learn it .

So how can we capitalize on empathy in the workplace ? Empathy isn't just a feel-good tool for your family and friends . Research shows that when managers and employers are good listeners , employees are more loyal , engaged and productive . That means these same chatbot skills of paraphrasing , validating and following up are powerful at the office too . But let's be real .

Nobody wants to feel like they're talking to a robot . It's about weaving these skills into a genuine human interaction we all want to know that we're listened to , but we also want to sense that that other person truly cares . So what are the limits of AI empathy by

now , you might wonder ? Are chatbots set to replace human listeners entirely ? Not quite . There are a couple important caveats . First , repetition and formulaic responses . Chatbots can begin to feel a bit stale or repetitive over longer conversations . A single chat might feel supportive , but after multiple interactions you might crave more nuanced , spontaneous understanding .

So don't be that Technical glitches and hallucinations . Chatbots sometimes respond with misinformation or stray off topic Actually a lot like some of the people I know and emotional depth . Real people can share experiences , emotions and vulnerability in ways AI cannot replicate .

Chatbots don't have genuine feelings , so their empathy , while helpful , can feel limited or performative . In fact , most people still prefer waiting to talk with an actual person over an immediate conversation with a bot Think calling customer service but especially for emotional matters .

That's because there's something uniquely beautiful about human empathy it's coming from someone with their own limitations , someone who might be tired or stressed themselves but chooses to offer comfort anyway , and that shared humanity is precious . So what are the key takeaways we can

learn from chatbots ? Well , if there's a central lesson in all of this ? It's that chatbots aren't empathic because they possess some kind of magical emotional intelligence . They're empathic because they avoid the big mistakes we humans tend to make when we're trying to help others .

Help others , and if we can learn to dodge those pitfalls , focusing more on the other person , validating their experience and asking thoughtful follow-ups , then we can bring a little chatbot magic into our own human conversations . It's not about becoming robotic on how we talk to people .

It's about listening first , acknowledging the other person's feelings and only then deciding together on next steps . So that wraps up today's conversation on empathy and what chatbots can teach us about becoming better listeners .

Thank you to Jamil Zaki for his insightful article and if you're interested in learning more about his work , you can check out his latest book Hope for Cynics the Surprising Science of Human Goodness . And thank you for joining me today . I hope this episode inspires you to try out a few of these chatbot-inspired listening skills in your next conversation .

And remember great listening isn't just about the words we say . It's about showing others that we truly care . I'm Sam Rhee and you've been listening to Botox and Burpees , and if you enjoyed this episode , please subscribe and share with a friend who could benefit from better listening skills . Take care , stay curious and keep listening . Thank you .