When the chips are down

As I settled in to enjoy my weekend last Friday, you might well imagine my reaction to the headline, “Armed police handcuff teen after AI mistakes crisp packet for gun in U.S.”

It would probably have been something along the lines of, “Here we go again,” except that I had just fallen ill so it was more like, “Blergh,” and I most certainly did not enjoy my weekend. I did, however, have a lot of time to reflect on the incident, which saw a false positive resulting in multiple officers, guns drawn, detaining an innocent student.

Now, I have a lot of fingers to point here. I don’t think anyone came out of this event looking particularly good. Not the security company, which washed its hands of the incident when it felt it had performed its due diligence; not the school district, which determined that it was, indeed, NOT a gun before telling its school resource officer; and not the police, who decided that it somehow still required this level of response.

With the further integration of security technology into the public safety apparatus, it’s important that we as an industry get things right. Seconds matter in a potential school shooting so I can’t fault tech provider Omnilert for its response in this crisis, per se. I think they made the best call in the situation given, but with so many moving parts in an environment, it’s important to consider the consequences that could transpire. This was one jumpy response away from a tragedy.

AI-facial and gun detection technology is impressive and terrifying in equal measure, precisely because of its potential. While Omnilert does call gun detection “messy,” it’s for that reason that we need to assess whether a technology is ready for the purpose it’s put toward and whether it can be all that it claims to be.

Or if it’s a bag of chips.