Eye On AI: Regulation For Artificial Intelligence Is Coming, Perhaps One Industry At A Time

This column is a look back at the week that was in AI. Read the previous one here.

Perhaps the only thing rising faster than the valuations of generative AI companies like OpenAI and Anthropic is the talk of regulation around the technology.

We already saw the likes of Sam Altman on Capitol Hill talking about regulation in the space, and Europe seems a couple of steps ahead of the U.S. when it comes to the topic.

This week, regulation talk sprung up again, not by an elected politician, per se, but rather from Securities and Exchange Commission Chair Gary Gensler — who never seems to shy away from talk of regulation.

In an interview with the Financial Times, Gensler made the case that regulators must figure out how to manage the financial risk posed by AI platforms used by financial institutions.

He said a financial crisis triggered by AI is “nearly unavoidable” within a decade without regulation. His concern stems from too many financial institutions using the same AI base model or aggregator, and the fact that model could sit outside the purview of regulators at some big tech company.

What’s interesting is that so much talk about regulating AI seems to be very big picture and ambiguous — usually by politicians — and oftentimes comes down to a sci-fi question like “will it take over the world?”

Serious talk about regulation by industry is likely what’s needed (although even Gensler admitted what is needed in the market is almost too “horizontal” for regulators as it affects more than just certain funds or brokers).

That undoubtedly will affect startups in those industries. There are many fintech and analytic companies trumpeting their use of AI. As talk of regulation by the SEC gets more serious, those descriptions may be toned down, as will the multiples those startups were hoping to see.

Of course, this should not be news for seasoned founders and entrepreneurs. In the past, tech startups have often complained about the difficulties of getting into highly regulated sectors such as finance, law and health care. Still, now it seems like some of those same industries have become early adopters when it comes to AI to some extent or another (as we talked about here, in the case of legal tech).

Perhaps now the SEC will make banks, asset managers and others turn back to that simpler time.

Things that caught our eye and other stuff:

• AI certainly makes it tough to tell what is real and what isn’t. Well, New York-based Reality Defender is looking to help, as the startup raised a $15 million Series A this week led by DCVC. The company is one of several startups developing tools to detect deepfakes and other AI-generated content across audio, video, images and text. Reality Defender’s text detection tool enables clients to scan a document and see color-coded paragraphs of AI-generated text.
• Another startup to raise cash in the last week or so was Germany-based Orbem. The Munich company raised a Series A worth almost $32 million led by 83North. The startup is using AI technology in MRIs to drive efficiency. Using AI-powered MRI tech, the company can already scan chicken eggs to determine the sex in a contactless and noninvasive manner — while delivering results in just one second per egg. The idea is the technology takes on the roles of technicians and radiologists — making the MRI process faster and more viable. Perhaps AI will one day tell us if the eggs in our fridge are still good or not?

Illustration: Dom Guzman