[bevacqua]

"A computer can never be held accountable, so has increasingly been used to make management decisions." — IBM slide from 1979.

AI didn't wrongly convict anyone. People leveraging AI did.

[adhesive_wombat]

I believe the 1979 original read "...Therefore a computer must never make a management decision".

https://constelisvoss.com/en-gb/pages/a-computer-can-never-b...

[bmitc]

I think I actually like the misquote, as it's more powerful, although the original is more prescriptive.

[IG_Semmelweiss]

Can we have an honest conversation about all the people enabling X , not just those making management decisions?

If we design an algorithm that will be used to put people in jail by algorithm [1], is that exempt from culpability? What about the engineer that implements it? If they are not responsible, but the code is buggy and the bugs cause real harm, are they now to blame?

Where is the line that we should not cross ?

[1] https://www.cbsnews.com/news/hertz-to-pay-168-million-in-bog...

[mikaoj85]

That’s not quite the correct quote.

Source: https://twitter.com/MIT_CSAIL/status/1604884273789603842

[pclmulqdq]

This is exactly why computers are being used to make management decisions. Diffusion of responsibility is a hell of a drug.

[paxys]

And AI didn't cause his rape, the prison system did.

[pixl97]

Eh, an abusive system inside of an abusive system. It's best to look at it in its entirety to get a view of the systemic abuse.

[aredox]

Do you have a link to a source?

[tiborsaas]

I've found this: https://twitter.com/MIT_CSAIL/status/1604884273789603842

[polonbike]

That one slide can easily be found via google, but I cannot find the full presentation