President Biden to Instruct Federal Agencies to Protect A.I. from Foreign Enemies

­Arguably the biggest foreign threat isn’t from guns or bombs, but from cyber intrusions. And that digital menace will only get scarier with the onset of A.I. So it’s none-too-surprising that President Biden recently instructed intelligence and national security agencies to protect AI from foreign adversaries.

In typical federal fashion, Biden is about to sign a memorandum urging certain government agencies to be on the lookout for foreign cyber attacks via A.I. So, you know, nothing binding or concrete, though securing domestic A.I. systems should be (and in many cases, is) a top priority.

According to the NY Times, the memorandum states that humans should be “kept in the loop” with weapons-related A.I., and A.I. also shouldn’t be used for making decisions on “granting asylum, tracking someone based on their ethnicity or religion, or classifying a person as a “known terrorist” without human review.”

Curiously, this non-binding memorandum didn’t outright ban A.I. from making life-or-death decisions on the battlefield, with the guidance intentionally left vague. Does “in the loop” mean humans must sign off on offensive actions or that they be “informed” at some point in the process? After a hypothetical attack, mind you.

What’s even scarier is that, according to a recent survey, nearly half of all cybersecurity professionals are being excluded from the development, onboarding, and implementation of AI solutions.

One could easily surmise that cybersecurity teams should be amongst the very first people involved in A.I. development, but we may not learn that lesson until it’s too late.