Surveillance - and Control
Be careful
You have all seen it on television - the police, using surveillance cameras, catch a thief walking into a store and using face recognition are able to nab him (and it usually is a him, not a her). Success, correct?
Do you know that the average American, in 2020, was recorded by security cameras 238 times each week? And that was documented five years ago - the number has certainly increased since then.
Facial recognition - a great idea, correct? In a recent report the Innocence Project, devoted to release innocent people from incarceration, reported on the many times that facial recognition got it wrong. Unregulated and untested AI facial recognition technologies put innocent people at risk. It probably comes as no surprise, but is shocking nonetheless, that Black and Brown people are more often than whites to be incorrectly identified.
Finally, a recent article in the Bulletin of the Atomic Scientists published an article that should alarm everybody. It reported that on March 6, 2025, Axios reported on a new State Department social media program called “Catch and Revoke.” This program would use AI to review “tens of thousands of student visa holders’ social media footprints to find evidence of alleged terrorist sympathies.” Wow! What happened to the First Amendment? That is only one example of the use - or more accurately, the misuse - of data collected by AI to predict behavior.
In order for AI to be safely used in the many positive ways that are possible, and in order to be able to trust it, we need strict and firm regulation. The Food and Drug Administration certifies that the medicines that we buy and are prescribed for us are safe and effective. We trust them, and by and large that trust is justified, despite some rare examples of unsafe drugs that slipped through. We need a similar agency with very strict rules to make sure this technology is safe.
We have got to do better. We can do better. The advances in technology of all kinds are terrific, but we must make sure they are used safely and carefully, using well tested techniques.
As always, comments very welcome. And if you think this issue is important, spread the word.
My best, Alan