High school student surrounded by police and handcuffed after AI mistakes his bag of Doritos for a gun

[Stiqy]

"The program is based on human verification and in this case the program did what it was supposed to do which was to signal an alert and for humans to take a look to find out if there was cause for concern in that moment," Rogers said.

Maybe the human verification should be reviewing the camera footage before dispatching 8 police cars.
I mean, I presume the footage must be *right there* in the cloud, attached to the AI alert, able to be forwarded to the commanding officer or anyone to have a 30 second look before dispatching.

 

[StrikerRocket]

"Put your doritos down right now!"

AI makes a lot of money while making so much mistakes.

A.I does not make a lot of money at the moment, it's only the beginning, and the financial AI bubble might explode sometime... so much investment and so little return for now. I may be wrong, but some are going to lose something in the process when all this tech sector gets a cleanup.

(as a side note: write "so *many* mistakes", not "so *much* mistakes". Much is used in front of uncountable nouns such as water in "Don't drink so much water", or butter "Don't eat so much butter!" when talking about a global quantity, not discrete elements. On the other hand, "don't eat so many biscuits!" ;-) Don't get mad at me, I teach English, so, this is just a profesionnal quirk...

 

[jeffbert]

Good thing it was not a bag of FRITOS!

 

[Megalomaniac]

When you're certain skin tones, your wallet gets mistaken for a gun by the human eye.

 

[Megalomaniac]

When you're certain skin tones, your 3 Musketeers bar gets mistaken for a gun by the human eye.

 

[Megalomaniac]

When you're certain skin tones, your cellphone gets mistaken for a gun by the human eye.

 

[PierreJG]

I just don't understand how they can try and justify this. A firearm and a crankled up packet of chips won't even have the same shape. That is the main problem I have with AI. It is made by humans and humans make mistakes. As an Web developer I have to use AI but I try to use it minimal.

 

[Liljonny11]

Absolutely crazy that they don't even review the images before pulling guns on people. Not too long before we got Cops blaming Ai for shooting kids.

 

[Liljonny11]

Well, firstly, the reviewer is likely another cop, someone who is trained to see a threat in everything. The AI did the equivalent of waiving a red flag in front of a bull, and we wonder why the bull decided to charge? Secondly, who wants to be the person that has to admit 'the AI flagged a threat and I discounted it, and now kids are dead'? With all the shootings and criticism police have had in the past decade for failing to stop school shooters quick enough, there is no way a cop would respond any other way then they way they did.

While the reviewer erred, the whole system has contributed to make such a mistake a near certainty. And that is the problem. So now we have the situation where having 8 armed police pull guns on an innocent teenager makes schools 'safer' because 'the system worked as intended'.

Yay, I am so glad we got here as a society, go us...

except a different cop reviewed the footage and realized what happened. The first cop should have come to the same realization.

 

[ScottSoapbox]

That's a really easy solution stop voting against school choice bills we need to put it into the public schools anyway in the absolute cancer of what the ideology they keep spouting

I mean I vote for school choice, but politicians are more swayed by teaching union donations than losing my support.

 

[thbaseer]

Although I'm late to seeing this news and I'm sure others elsewhere have already pointed it out---I'll bet a million dollars that Omnilert's AI factored in the kid's skin color/ethnic appearance in determining that the unidentified object (bag of chips) in hand had a chance of being a gun.

Demographics data has always been inaccurate, especially historic/legacy data, due to the means of reporting and collection and recordkeeping of such data. AI systems are tapping into these databases and are essentially ingrained to be systemically racist.

Data needs a reset.