Comparing Physician and Artificial Intelligence Chatbot Responses to Patient Questions Posted to a Public Social Media Forum

RT @denisjabaudon: Patients find ChatGPT's answers more empathetic than actual physicians'. Sounds like a headline from The Onion but no,…

@AndreasThinks this might be of interest https://t.co/KS4majKTDU

RT @denisjabaudon: Patients find ChatGPT's answers more empathetic than actual physicians'. Sounds like a headline from The Onion but no,…

RT @denisjabaudon: Patients find ChatGPT's answers more empathetic than actual physicians'. Sounds like a headline from The Onion but no,…

RT @denisjabaudon: Patients find ChatGPT's answers more empathetic than actual physicians'. Sounds like a headline from The Onion but no,…

RT @denisjabaudon: Patients find ChatGPT's answers more empathetic than actual physicians'. Sounds like a headline from The Onion but no,…

RT @denisjabaudon: Patients find ChatGPT's answers more empathetic than actual physicians'. Sounds like a headline from The Onion but no,…

RT @denisjabaudon: Patients find ChatGPT's answers more empathetic than actual physicians'. Sounds like a headline from The Onion but no,…

RT @denisjabaudon: Patients find ChatGPT's answers more empathetic than actual physicians'. Sounds like a headline from The Onion but no,…

Patients find ChatGPT's answers more empathetic than actual physicians'. Sounds like a headline from The Onion but no, real stuff... https://t.co/HEcvaxHYcf https://t.co/od8uhiDW1B

@BadMedicalTakes For some questions, AI is already better than human doctors: https://t.co/FVVKidPoA8

RT @jonc101x: @FutureJurvetson @stefan_bod @StanfordMed @LMWilliams_PhD There have been a couple high profile studies in past year showing…

RT @MachineElvesApe: @Cephii1 @sdixitmd @bindureddy https://t.co/66F4377Hmc Al responses were found to be 3.6 times higher in quality and…

Revolutionary times ahead https://t.co/fvIvzw6QjF

@Cephii1 @sdixitmd @bindureddy https://t.co/66F4377Hmc Al responses were found to be 3.6 times higher in quality and nearly 10 times more empathetic than those of physicians. This was GPT-4...Only going to get better.

@chburkart @sdixitmd @bindureddy https://t.co/66F4377Hmc

https://t.co/Pj94jqa7Qk Y'all got smoked by GPT-4. That argument of 'AI hallucinates' is so silly. All that means is that it is confidently wrong sometimes. Know who else is confidently wrong sometimes? Humans! It's a matter of how often the AI is wrong c

@guycalledfrank Yes there are ethical & legal challenges, but humans are unreliable too, and not particularly accountable. https://t.co/a7FJr2e0wv https://t.co/i5V5gFfvty. https://t.co/fvaAxRQwu2

RT @jonc101x: @FutureJurvetson @stefan_bod @StanfordMed @LMWilliams_PhD There have been a couple high profile studies in past year showing…

@FutureJurvetson @stefan_bod @StanfordMed @LMWilliams_PhD There have been a couple high profile studies in past year showing bot-generated responses to patient's medical questions were rated higher in quality AND in empathy compared to doctor-written respo

Stand even further away from a patient? Is that the takeaway? 😑 https://t.co/yCPqlTXdAw

@cyber_gilgamesh @DeryaTR_ https://t.co/77uZ5rMsWn

@LionBlogosphere People find chatbots more empathetic than doctors when compared head-to-head. And that was a year ago, when chatbots were less sophisticated. https://t.co/2agtPNKOTn

@StanfordMed @StanfordMedMuse @LisaRotenstein @StanfordWellMD @RoseLikeTheFlwr @StrehlowMatthew @StanfordEMED @LalehG_MD How can #AI help the clinician experience? https://t.co/spbvZW4PQ3 @JohnWAyersPhD https://t.co/JyopqUvSHr

Hola @bilinkis, ayer publicaste un reel referente a este paper: https://t.co/CFvgHevYDz El cual fui a leer directo luego de ver el reel, justamente porque pega directo en la disciplina que estoy estudiando y la ciencia que estoy aprendiendo

@FaithandSparkl1 You can find the study here: https://t.co/2Vq1iicfD0. I'll probably write a longer article about it on Medium some time soon!

RT @EricTopol: Noteworthy: a new study comparing #ChatGPT vs doctors for responding to patient queries demonstrated clearcut superiority of…

@dem_taz @EnfrmraSaturada https://t.co/JIp4jwGj2g. Más reflexion necesitamos.

RT @EricTopol: Noteworthy: a new study comparing #ChatGPT vs doctors for responding to patient queries demonstrated clearcut superiority of…

@ColinDavdButler Very good article. But reason 6 doesn't deal with the fact that pple might be brought to the point by our neolib gvt where they prefer a chatbot to a human working in an increasingly underresourced system eg general practice https://t.co/f

The chatbot responses were preferred over physician responses and rated significantly higher for both quality and empathy." https://t.co/iROjeZ5hUn As @EricTopol can attest there are a lot of papers like this.

AI is becoming more empathetic than humans https://t.co/PZxDc09TJx

RT @JeremyNguyenPhD: 2/ Link to paper. - people found AI responses more empathetic than human doctors: https://t.co/1JjWkRLohc

2/ Link to paper. - people found AI responses more empathetic than human doctors: https://t.co/1JjWkRLohc

RT @Empathy_Centre: 📰 NEWS: A team of researchers from the University of California San Diego has suggested that artificial intelligence co…

RT @Empathy_Centre: 📰 NEWS: A team of researchers from the University of California San Diego has suggested that artificial intelligence co…

RT @Empathy_Centre: 📰 NEWS: A team of researchers from the University of California San Diego has suggested that artificial intelligence co…

📰 NEWS: A team of researchers from the University of California San Diego has suggested that artificial intelligence could make healthcare more empathic 🔗 Read more here: https://t.co/X4Kd8U95Jf #empathy #medicaleducation https://t.co/TJLYfWZxW1

RT @LouFerreira9: Comparing Physician and Chatbot Responses to Patient Questions https://t.co/rJOYMk7sAY

Comparing Physician and Chatbot Responses to Patient Questions https://t.co/rJOYMk7sAY

"Söz eyleminin performatif/edimsel yanı o beyanın pragmatik ‘etkisi’yle aynı şey değildir." https://t.co/5LBN1yRENG https://t.co/vNfXh5lknc

RT @MustBeMistry: 👀 GPT4 is rated higher in perceived empathy than some clinicians for text based communication. Lots of caveats! More que…

👀 GPT4 is rated higher in perceived empathy than some clinicians for text based communication. Lots of caveats! More questions than answers on this right now. https://t.co/zNq82uzTgw https://t.co/QKdvk8dIjv

RT @EricTopol: Noteworthy: a new study comparing #ChatGPT vs doctors for responding to patient queries demonstrated clearcut superiority of…

@nathaliaccosta1 There's more and more evidence coming out in favor of AI vs human HCP's in lots of areas. https://t.co/dnatt7dr2o Almost 10 times higher prevalence of empathy compared to physicians, and preference for AI responses in 78% of cases. Smal

■ Ref info: Ayers et al. "Comparing physician and artificial intelligence chatbot responses to patient questions posted to a public social media forum." JAMA internal medicine 183.6 (2023): 589-596. ■ Ref link: https://t.co/IaLNt5gjJV ■ note link: https://

@davidmcw @BBC @Columbia @dalkeybookfest Way off the mark with your prediction on data analysts. This job isn’t going away as these are the ppl who can use the tech. Also incorrect about EQ prediction - LLMs already showing better EQ than medical professio

@emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple controlled studies. JAMA study: https://t.co/W3X0eHubF4 Harvard study: https://t.co/w02kU1K3oK https://t.co/bqUCSolFWa

@emollick: @Grady_Booch I don’t love the insults, Gary, but this is just the state of the research. GPT-4 is consistently rated as higher in apparent empathy than humans in multiple controlled studies. JAMA study: https://t.co/W3X0eHubF4 Harvard study: …

(via https://t.co/FJ5lJIH59a via https://t.co/foWXEqy1oT )

@codewithdev Here's a link to the study: https://t.co/Fb0QDCglrk

@codewithdev Thanks, Dev. It's not my study, but yep, I'll see if the full data is available. The study is published here: https://t.co/Fb0QDCglrk

@leongreen This is the study, Leon: https://t.co/Fb0QDCglrk

@TonaWrites Thanks Tona. This is the source study: https://t.co/15TtfYXlCr

@emollick: Empathy and quality of answers on reddit about common medical issues, doctors vs. GPT-3.5. https://t.co/W3X0eHtDPw https://t.co/KcBHT2xvnY

RT @jayvanbavel: I had to read this twice: ChatGPT is not only more accurate than medical doctors but also *more* empathic.

@aryehazan We’re already seeing this in medicine, though I’m not sure how highly we should rate Dr’s verbal skills. https://t.co/Q6Ft4ae7oA

RT @ThomasScialom: RHLF versus immitation learning explained in one tweet

RT @emollick: Empathy and quality of answers on reddit about common medical issues, doctors vs. GPT-3.5. https://t.co/BttWRfaQ7B https://t.…

RT @emollick: @Grady_Booch I don’t love the insults, Gary, but this is just the state of the research. GPT-4 is consistently rated as highe…

RT @IntuitMachine: Interesting, GPT-4 appears to have "artificial empathy."

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

RT @matspike: putting "rates highly in apparent empathy' on my LinkedIn bio

Not surprised.

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

putting "rates highly in apparent empathy' on my LinkedIn bio

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

RT @emollick: @ChombaBupe Of course AI apparent empathy is merely masquerading as empathy. AI does not have real empathy. But I am not sure…

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

تحقیقات جدید نشون میده که: مدل GPT-4o در در ابراز همدلی با بیماران عملکرد بهتری نسبت به پزشکان داشته! این نتایج بیانگر اینکه مدل های جدید AI مراجعه کنندگان رو بهتر درک میکنن و پاسخ هایی متناسب با روحیه و حال روز اونها ارائه میدن. https://t.co/5plfzlnO6g

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

RT @IntuitMachine: Interesting, GPT-4 appears to have "artificial empathy."

LoL

RT @IntuitMachine: Interesting, GPT-4 appears to have "artificial empathy."

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

They have about as much empathy as a politician trying to get elected. There is nothing in the system that allows for this. Chatbots are just using crowdsourced replies!

People are missing the point here The chatbots are not doing as well as humans, they are doing *much better* than humans For example, in the JAMA study, a chatbot had 3.6x better odds of giving a quality answer, and 9.8x odds of giving an empathetic answ

RT @IntuitMachine: Interesting, GPT-4 appears to have "artificial empathy."

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

Interesting, GPT-4 appears to have "artificial empathy."

@ChombaBupe Of course AI apparent empathy is merely masquerading as empathy. AI does not have real empathy. But I am not sure why measuring other people’s perceptions of empathy is nonsense, especially when apparent empathy matters in the world in many cas

Not only does Ethan have zero clue about empathy, he also doesn't understand the word 'apparent'. Ethan needs a dictionary and a course in critical thinking.

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

RT @emollick: @Grady_Booch I don’t love the insults, Gary, but this is just the state of the research. GPT-4 is consistently rated as highe…

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

RT @emollick: Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple c…

Since people aren’t reading the post, yes,GPT-4 is consistently rated as higher in apparent empathy than humans in multiple controlled studies. JAMA study: https://t.co/BttWRfbnX9 Harvard study: https://t.co/85gVxkjIGc https://t.co/ugzhUwY352

@Grady_Booch @emollick I think he specified *apparent* empathy intentionally. And this does seem to have been demonstrated in at least one interesting case https://t.co/QTn8PgqO7B

@Grady_Booch I don’t love the insults, Gary, but this is just the state of the research. GPT-4 is consistently rated as higher in apparent empathy than humans in multiple controlled studies. JAMA study: https://t.co/BttWRfbnX9 Harvard study: https://t.co/

I guess doctors are missing the "You are polite, emphatic, and helpful virtual physician who... " part of the prompt?

RT @jayvanbavel: I had to read this twice: ChatGPT is not only more accurate than medical doctors but also *more* empathic.

RT @jayvanbavel: I had to read this twice: ChatGPT is not only more accurate than medical doctors but also *more* empathic.

Impressive, however it only covers common medical issues, I would imagine the model to generalize much more poorly. Physicians know about rare diseases and can probably identify those with ease, whereas recall on chatbot is probably really low for rarer d

People overestimate themselves and underestimate AI - and this is just 3.5!