3 ms·
Things like this really favor models offered from countries that have fewer legal restrictions. I just don't think it's realistic to expect people not to have a
by sfblah 11mo ago
Things like this really favor models offered from countries that have fewer legal restrictions. I just don't think it's realistic to expect people not to have access to these capabilities.
It would be reasonable to add a disclaimer. But as things stand I think it's fair to consider talking to ChatGPT to be the same as talking to a random person on the street, meaning normal free-speech protections would apply.
- thinkingtoilet 11mo agoWhat capabilities? The article says the study found it was entirely correct 31% of the time.
- Jweb_Guru 11mo agoThis is what really scares me about people using AI. It will confidently hallucinate studies and quotes that have absolutely no basis in reality, and even in your own field you're not going to know whether what it's saying is real or not without following up on absolutely every assertion. But people are happy to completely buy its diagnoses of rare medical conditions based on what, exactly?
- simianwords 11mo agoGive a single example using gpt-5 thinking.
- Jweb_Guru 11mo agoGPT-5 thinking is one of the biggest offenders and it's quite incredible to me that you think it doesn't hallucinate. It makes me strongly suspect your own judgment is impaired. Also, what do you mean asking for "reproducible examples"? Is it somehow not a valid example if it only sometimes makes up citations?
- simianwords 11mo agoShare any example where I can get it to make an obvious error
- ipaddr 11mo agoDoes it say how often doctors are correct as a baseline?
- matwood 11mo agoThe study is more positive than the 31% conveys. https://www.ctvnews.ca/health/article/self-diagnosing-with-ai-canadian-study-finds-chatgpt-can-get-it-wrong/ https://www.ctvnews.ca/health/article/self-diagnosing-with-a... The researchers compared ChatGPT-4 with its earlier 3.5 version and found significant improvements, but not enough. In one example, the chatbot confidently diagnosed a patient’s rash as a reaction to laundry detergent. In reality, it was caused by latex gloves — a key detail missed by the AI, which had been told the patient studied mortuary science and used gloves. ... While the researchers note ChatGPT did not get any of the answers spectacularly wrong, they have some simple advice. “When you do get a response be sure to validate that response,” said Zada. Which should be standard advice in most situations.
- sarchertech 11mo ago>I think it's fair to consider talking to ChatGPT to be the same as talking to a random person on the street That’s not how companies market AI though. And the models themselves tend to present their answers in a highly confident manner. Without explicit disclaimers, a reasonable person could easily believe that ChatGPT is an authority in the law or medicine. That’s what moves the needle over to practicing the law/medicine without a license.