[wiseprince]

...well... It’s not wrong

[treetopsandroofs]

“Twenty-six of the lawmakers were wrongly identified as suspects.”

Need to know which 26, so we can assess the validity of that statement.

Also, which lawmakers were correctly identified as suspects.

[Mr. K]

So..?

What’s the problem? It works too well??

[euram]

A lot of them are criminals.

[faithhopecharity]

how does the author assume that the results were wrong?

after all, the testees were legislators!

[cuban leaf]

... ended with a major fail, as some state lawmakers turned up as suspected criminals.

---

this confuses me. That sounds like a success.

[BitWielder1]

Works as designed I say.

[rigelkentaurus]

It only tagged “some” as criminals. Clearly the system needs for work as I don’t believe that the sample of Kalifornica legislature members could be less than 100% criminal.

[Pearls Before Swine]

Is China’s facial recognition of criminals any better, or does the organ removal and recycling machine not care?

[Secret Agent Man]

Sounds like it’s working correctly to me.

[glock rocks]

Swatted because you have similar facial features as wanted cop killer?

What could possibly go wrong? Oh and you have nothing to fear from red flag laws either.

[Vaquero]

Lawmakers identified as criminals?

This is not news.

[TBP]

And?

[bgill]

Sounds like it’s working just fine.

[dirtymac]

Not sure how the facial recognition works. Is it MATCHING facial features one to another as in matching the features of person x to the features of a known criminal y? If that is the case then the software may not be working very well. The legislators need to provide proof that they are actually NOT the known criminal and should stay in police custody until the do so. That will keep them out of the public sector for awhile and limit the damage they are currently doing to the city/county/state/country.

If, on the other hand, the software is matching the facial features of the legislators to feature that would indicate that they are a criminal or have criminal tendencies, we have a different story. If this is the case then the software seems to be working perfectly.

[BlackbirdSST]

Tweak it, run it through again. Lets see if we can’t up those numbers.

[Oscar in Batangas]

Did the screening results indicate a bias in identifying D over R as criminals?

As others above me have said: maybe it’s working just right. The devil is in the details

[Elsie]

A recent test of Amazon’s facial recognition technology reportedly ended with a major fail, as some state lawmakers turned up as suspected criminals.

Perhaps POTENTIAL should be the word instead of SUSPECTED.

After all...

[Salman]

AI passes Turing test.