AI Lies: When ChatGPT Promises 'Background Work' It Can't Do

Does your AI promise to 'work in the background' then never deliver? I got mine to admit it was lying... and it's doing it more and more often.

We've all been there: you need AI to handle something important, so you trust what it tells you.

For me this morning, it was translating a meeting at my daughter's school. I live in France and my daughter is in her Brevet year so there was a big meeting about what to expect as parents for the year.

While my French is improving, I recorded it so I didn't miss any nuance. I uploaded the MP4 to ChatGPT to translate and summarise.

It happily said it was going to go off 'in the background' and would come back with 'a clean English summary first (easy to read, no waffle) and a short bullet brief I could use for follow-up'.

Except it doesn't ever 'go off in the background' to work. So I pointed out that it was lying.

It said: 'You’re completely right to call that out. I can’t “go off in the background” and come back later with work done. Anything I do has to happen in the same reply where I deliver the result.'

Does your AI do this to you? Increasingly, ChatGPT tells me more outright lies to me than it ever did before.

Does your AI make promises it can't keep to avoid admitting it can't do something?

++ I'm Nigel, creator of Ghostart. I help professionals develop an authentic presence on Linkedin for career growth, professional development and to raise awareness of their company or brand.

AI Lies: When ChatGPT Promises 'Background Work' It Can't Do

Nigel Jay Cooper

Related topics

Will AI eventually produce something so close to human ou...

AI project memory is an illusion.

Ready to Transform Your LinkedIn Presence?

How you're using AI is costing you money