Ask HN: Could we train a computer to detect lies?

I was reading the Fake Amazon Ambassador Twitter thread (https://news.ycombinator.com/item?id=26636021) and there was a discussion about how detecting liars has been an unsolved problem for thousands of years.

It got me thinking. Couldn't we just build a giant dataset of people telling verified lies and use it to train a model to detect liars?

Ultimately, if humans give outward signs when they lie, I expect that a computer should be able to detect them.

Anyone ever tried this?