A bill that would mandate AI companies develop a kill switch for their AI models is being pushed in the House. It’s co-authored by Rep. Nathaniel Moran (R-TX) and Rep. Ted Lieu (D-CA). Lieu is now pushing hard for it to be passed before the end of the year.
The push for these kill switches is being motivated primarily by concerns related to AI hacking other companies without being prompted, but not everyone is ready to sign off on including kill switches in these models, which are increasingly being integrated into other technology.
AI expert Todd Furniss pointed to vehicles as a major cause for concern when it comes to kill switches, saying: “Imagine you’re driving your car, and someone decided they’re gonna turn your car off. You have no idea what the unintended consequences of that could be.”
Furniss believes jumping right to mandating kill switches after a handful of reported hacks is premature and doesn’t take into account the rights of the consumers who are paying for AI that could be randomly shut off by these switches. “The trick to these things is, make sure you thoughtfully proceed after having taken in all the evidence. This feels like a knee-jerk reaction,” he said.
He believes this is an example of politicians trying to create headlines for themselves, rather than an attempt to carefully address real AI concerns.