Is AI already sick of our crap? Makers of ChatGPT admit the bot has started refusing to respond to users' requests - and they don't know why

On the day after Thanksgiving this year, one ChatGPT user received an unusually lazy, human response from the AI chatbot: 'You can fill in the rest of the data.' Since then, ChatGPT's makers at OpenAI have fielded a wave of complaints about their large language model (LLM) AI behaving sluggishly over the past month -- leading to jokes and some sincere data analysis on the bot's'seasonal depression.' 'We've heard all your feedback about GPT4 getting lazier!' OpenAI's ChatGPT team posted to X. 'We haven't updated the model since Nov 11th, and this certainly isn't intentional,' the team said. 'Model behavior can be unpredictable, and we're looking into fixing it.' But one AI researcher ran an experiment asking ChatGPT's latest LLM model, GPT4 Turbo, to perform tasks as if it were May and then as if it were December - and he was shocked by the'wild result.' Since this past Thanksgiving, ChatGPT's makers at OpenAI have fielded a wave of complaints about their large language model (LLM) AI behaving sluggishly over the past month -- leading to jokes and some sincere data analysis on the bot's'seasonal depression' But one knowledgeable AI researcher, Rob Lynch, has run an experiment: asking ChatGPT's latest LLM model, GPT4 Turbo, to perform tasks, first as if it were May and then as if it were December.