Give OpenAI CEO Sam Altman credit, because he said all the right things during his congressional testimony on Monday — telling lawmakers, among other things, that his research lab, which launched the AI-based chatbot ChatGPT, absolutely favors federal regulation. And that tech leaders like him are very much in favor of government oversight to help ensure the industry doesn’t usher in some dark, dystopian chapter of humanity via the emergence of advanced AI systems.
And yet the more people talk about this — about federal regulation and oversight, specifically, as a kind of must-have and last line of defense, as it were, to protect us from the negative consequences of AI — the more I’m convinced that it’s basically a pipe dream.
I’ll grant you, maybe this kind of cynicism is a function of getting older. Having said that, though, when is the last time any of us has seen congressional leaders pull together and figure out how to address an emerging threat or paradigm-altering event in advance in a way that was actually productive and, you know, did something? Because I can think of a whole lot of examples of our august federal officials squandering one opportunity after another and mishandling easy layups of the sort that we pay them for.
Sam Altman congressional testimony
Nevertheless, that’s what tech leaders like Altman think we need to have. “We think that regulatory intervention by governments will be critical to mitigate the risks of increasingly powerful models,” Altman said during his opening remarks to the Senate Judiciary Subcommittee on Privacy, Technology, and the Law.
At one point during his appearance, Altman respond affirmatively to a Senator who wondered whether something like the Nuclear Regulatory Commission might be a model for how AI could be regulated. Other Senators pointed to the FDA as a potential model. Really? The same FDA responsible for the drug-approval process taking as much as 15 years, that progress-killing FDA?
Generally-speaking, the federal government is a morbidly obese behemoth where progress and good results go to die, a house of cards that’s unaccountable for its failures and whose successes are few but which nonetheless hoovers up tax revenue like water in the desert. You want to talk about a federal agency kicking into gear to protect us from a threat, how about just two years ago? We literally have a federal agency that has the words “disease control” in its name, but raise your hand if you think the CDC did an adequate job controlling the spread of the pandemic.
No, don’t give me Trump this or Biden that — I’m simply asking, did the agency tasked with protecting us from the very thing that caused a generational amount of pain, damage, and death do an adequate job of controlling the pandemic? Of course not. Or how about the REAL ID law that goes into effect in 2025, which means most of us need to get new driver’s licenses in order to be able to fly domestically, since they’ll have updated security measures associated with them?
That law was passed by Congress in 2005, in response to the 2001 terror attacks … 22 years ago. That’s the federal government you think will put guardrails in place to protect us all from AI run amok? To borrow a Biden-ism: Come on, man.
Opinion: Don’t look to Congress to save us from AI
I hope these end up being the ramblings of a cynic who’s ultimately proven wrong about all this, I really do. And obviously, I suppose, the federal government still does some things well. But for every positive example you throw at me, it’s frustratingly easy to come right back with how well it doesn’t work. Octogenerians who can’t even come together on something as simple as raising the debt ceiling in order to accommodate the government spending that they’ve already approved do not strike me as well suited to understand the ramifications of ChatGPT.
I personally think some of the fears about AI — specifically, regarding worker displacement and other dark worries about how the technology might change our world — might be a little overblown. But I’ll be surprised if we all make it to the other side of our brave new AI world unscathed, thanks largely or in part to help from Uncle Sam.