In the age of artificial intelligence, it seems even a bag of Cool Ranch Doritos can get you held at gunpoint.
That’s exactly what happened to 16-year-old Taki Allen outside Kenwood High School in Essex, Maryland, on October 20, after Baltimore County Public Schools’ (BCPS) AI-powered surveillance system mistook a crumpled chip bag in his pocket for a firearm. Within minutes, eight police cars swarmed the scene. Officers drew their weapons, ordered the teen to the ground, and handcuffed him — all because an algorithm panicked.
And yet, in the aftermath, officials claim the system “worked as designed.”
Welcome to the new normal — where “safety” means surrendering common sense to machines and bureaucrats who refuse to admit when technology goes too far.
When Artificial Intelligence Meets Administrative Incompetence
The AI system, Omnilert, is supposed to scan thousands of school security cameras for guns and alert human operators for verification. In this case, the algorithm flagged the shiny, oddly shaped chip bag as a possible firearm. Human reviewers saw it and — astonishingly — thought it was suspicious enough to send to school officials.
From there, a game of bureaucratic telephone ensued.
- The central security office cleared the alert.
- But the principal never got the memo.
- She called the school resource officer.
- Who called the police.
- Who arrived with guns drawn.
By the time they realized it was all a misunderstanding, Allen had already been handcuffed, searched, and humiliated in front of his peers.
The “human verification layer” everyone brags about? It failed. The supposed “checks and balances” of AI oversight turned into a multi-agency game of panic ping-pong.
Officials Defend the Indefensible
Superintendent Dr. Myriam Rogers held a press conference days later to assure parents that the system “worked as designed.”
Let that sink in:
A kid gets detained at gunpoint over a snack, and the superintendent says that’s how the system is supposed to work.
To her credit, Rogers did promise an internal review to “refine procedures.” But the tone was unmistakable — better safe than sorry, even if that means scaring innocent students half to death.
Principal Kate Smith, meanwhile, sent out a polished letter reaffirming the school’s “commitment to safety” and offering counseling to students. No apology. No accountability. Just more assurances that everyone followed protocol — a phrase that has become bureaucratic shorthand for “don’t blame us.”
The Real Problem: Outsourcing Judgment to Algorithms
The problem here isn’t just bad communication — it’s the blind faith in technology to do a job that requires discernment, empathy, and context.
We’ve let fear of school shootings drive us into the arms of “AI safety theater” — expensive systems that promise peace of mind but deliver false alarms, data collection, and racial disparities. The same politicians who lament the “school-to-prison pipeline” are quietly installing surveillance networks that treat every student like a potential suspect.
It’s no coincidence that Taki Allen is Black. AI systems trained on flawed datasets often misidentify darker skin tones and ordinary objects as threats — a well-documented bias in facial recognition and object detection technology.
If an algorithm can’t tell a chip bag from a gun, what happens next time it mistakes a phone, an umbrella, or a water bottle? Do we keep calling that “working as designed”?
The Safety Trap
No one disputes the need for safer schools. But there’s a point where “erring on the side of caution” becomes a cover for institutional cowardice.
When you have police drawing weapons on a teenager because an algorithm said so — and the superintendent calls it “the system working” — something has gone profoundly wrong.
This isn’t a failure of AI. It’s a failure of leadership.
The adults in charge are so terrified of being blamed for not acting that they’ll defend any overreaction, no matter how absurd, just to prove they “did something.”
A Lesson in Accountability
Councilmen Julian Jones and Izzy Patoka have called for a full review of Omnilert’s use in Baltimore County schools. Good. But let’s not allow the review to be another whitewashed report that “finds no wrongdoing.”
Someone needs to answer the hard questions:
- Who authorized AI systems that can trigger armed police responses from a blurry camera feed?
- Why wasn’t the “human verification” process actually verified?
- And who will be held responsible for the trauma inflicted on a 16-year-old who committed the crime of eating Doritos after practice?
The Bottom Line
This isn’t just about AI — it’s about a culture of fear-driven governance that hides behind buzzwords like “innovation” and “safety.” Maryland’s leaders have turned schools into surveillance labs, spending taxpayer money on systems that make parents feel secure while putting students at risk.
The promise of “smart security” was supposed to make us safer. Instead, it’s teaching a new generation that the government’s first instinct is to treat them as threats.
If a bag of Doritos can set off a full-blown police response, then maybe the real danger isn’t what’s in students’ pockets — it’s what’s in the policies of the people running the schools.
Support Independent Journalism
MDBayNews is funded solely through reader support. Help sustain investigations that hold public institutions accountable.
Donate, Buy Me a Coffee, or Subscribe for Updates.
Discover more from Maryland Bay News
Subscribe to get the latest posts sent to your email.