ChatGPT
-
LLMs … first six months of 2025 in review.
Link to video -
New studies show how generative AI has influenced human speech:
https://www.washingtonpost.com/opinions/2025/08/20/chatgpt-claude-chatbots-language/
Researchers found signs of human speech patterns becoming more like AI, and can trace back to certain patterns more prevalent in certain poorer countries where cheaper labor were hired to fine tune AI language models.
Then there is the counter movement where some people who are aware of AI influence to in turn deliberately moderate their own speech to sound less like AI.
-
New studies show how generative AI has influenced human speech:
https://www.washingtonpost.com/opinions/2025/08/20/chatgpt-claude-chatbots-language/
Researchers found signs of human speech patterns becoming more like AI, and can trace back to certain patterns more prevalent in certain poorer countries where cheaper labor were hired to fine tune AI language models.
Then there is the counter movement where some people who are aware of AI influence to in turn deliberately moderate their own speech to sound less like AI.
-
https://www.nytimes.com/2025/09/13/business/chatgpt-financial-advice.html
More people are turning to generative A.I. chatbots for financial advice, whether it’s for debt management, better saving strategies or stock picks.
-
My FIL just today told me he’s switching insurance companies after AI ran a comparison of his options out there.
@89th Did he just tell his location and age, etc. and what he was looking for?
-
-
https://www.washingtonpost.com/technology/2026/01/26/chatgpt-health-apple/
WaPo tech columnist have ChatGPT and other AI services access to his large trove of Apple Health data and clinical records asking the AI services to grade his overall health, and found that the answer varied wildly from query to query, even from the same AI services.
The takeaway message seems to be that AI is not ready for prime time when it comes to giving personalized healthcare advice.
-
Only as good as the stuff it is trained on. As @mik said in the other thread about training an AI only on 19 century data and asking it for medical cures.
