“Tesla recall, MSFT Bing fail, and Google Bard fail are NOT independent; each reflect the fact that you cannot build AI in the real world from Big Data and deep learning alone.

Too many edge cases and not enough reasoning. We need new approaches; current AI has been oversold.”

@garymarcus

From: https://twitter.com/GaryMarcus/status/1626320733029928960

Gary Marcus on Twitter

“No. Tesla recall, MSFT Bing fail, and Google Bard fail are NOT independent; each reflect the fact that you cannot build AI in the real world from Big Data and deep learning alone. Too many edge cases and not enough reasoning. We need new approaches; current AI has been oversold.”

Twitter
@bsletten @garymarcus zero understanding in the current ML parrot factories
@bsletten @garymarcus sounds a bit like "first snowflakes of the next AI winter" #groundhogday2023 ;) ... Agree with the "not enough reasoning" part, though!

@AxelPolleres @bsletten @garymarcus

One of the reasons that I'm skeptical about the US of AI in Emergency services, the very reason of the existence of emergency services are the edge cases.

@bsletten @garymarcus

I am not really worried about what AI/ML gets right or wrong. The risk is what authority people invest in it. As with any new technology, it isn't inherently good nor bad, it comes down to how it is used.

@Tbeauchamp I don’t disagree. I think all of the mentioned uses were inappropriate and all three tech leasing companies failed a most basic ethics test.

@bsletten

That is exactly the problem. Ethics should be one of the first considerations in any business decisions, not something that is deferred to handle if it becomes an issue