Don't blindly trust what AI tells you, says Google's Sundar Pichai

She also urged Google to take more responsibility over its AI products and their accuracy

Not really possible when the technology is built around probability. In other words, it is simply *not* deterministic.

It depends on how the dice roll. Ask the same question in a slightly different way or at different time and get a different answer.

All generative AI output should begin and end with "Trust this output as much as you would trust a 5-year-old to control your life".

And 4chan has (had? Haven't been there in a while) "only a fool would take the things here as fact"

Doesn't matter, a good chunk of people are going to take everything at face value

It's hard not to trust answers which are superhumanly plausible.

Imagine being offered a choice of two oracles. The first oracle would give you an answer that is inspiring yet mysterious and oblique. The second gives an answer that is guaranteed to be unoriginal but also highly, highly plausible. It also maintains a record of all the questions and makes them available to unknown persons.

> However, some experts say big tech firms such as Google should not be inviting users to fact-check their tools' output, but should focus instead on making their systems more reliable.

That is a very strange thing to say. They both are and should be doing both.

Should I know, I mistakenly trusted Google when it told me there was a live music event last Sunday in a local bar, turns out Google's AI got confused because there is another bar with the same name in a different city here in Colombia, so I dressed up and everything just to go to a bar that was closed that day.

Completely at odds with everything the company is actually doing. Shameful nonsense.

... too late

and don't blindly believe a bbc headline these days...