It would be foolish to ban generative AI tools
International Institute for Analytics Co-Founder Tom Davenport, who is also the President’s Distinguished Professor at Babson College, joins Yahoo Finance Live to discuss OpenAI CEO Sam Altman set to testify before Congress, AI regulation, and the future of AI.
Video Transcript
– The trendiest phrase on Wall Street is getting some attention on Capitol Hill this week. Congress looking to regulate the rapid advancements in artificial intelligence. But for right now, they’re still playing catchup. Some events on the docket in Washington this week include, first, OpenAI CEO Sam Altman, who will attend a closed dinner with a bipartisan group tonight. Then tomorrow, the CEO will testify before Congress for the very first time.
Joining us to discuss what to expect is Tom Davenport, cofounder of the International Institute for Analytics and the President’s Distinguished Professor of Information Technology and Management at Babson College. Thank you for joining me in this morning. So first want to start off by what your expectations are for the sort of reception that Sam Altman is going to get on the Hill.
TOM DAVENPORT: Some combination of hostility and ignorance, I would guess. Members of Congress are not known for their technical expertise in general, and yet they want to demonstrate that they are on top of this issue and are controlling big tech. So that will drive– that will drive the reception.
– So Tom, is it going to be similar as what we saw with the emergence of crypto, where you had a lot of lawmakers not familiar with the technology and the platforms themselves sort of trying to ask questions about something that’s still emerging, versus this time you have big tech with a much greater understanding here and regulators once again playing catchup?
TOM DAVENPORT: Yeah, you know, we did see this in the past with social media, and there were a lot of discussions, a lot of testimony by leaders of social media companies. Not much changed, and I would suspect that that would be the case with regard to these hearings, as well. It’s a really complex issue, and Congress doesn’t seem to deal terribly well with those complex issues.
– Indeed, because you might end up with sort of regulation by enforcement. I mean, a lot of people are wondering, should this be on a case-by-case basis? Should it be impact-based, the number of people who perhaps get exposed to it? Or even some of these just focusing on critical industries. What’s the approach that you think lawmakers are likely going to take?
TOM DAVENPORT: Well, your graphic is quite interesting because we should not be addressing this on a state-by-state basis, and we probably shouldn’t even be addressing it on a country-by-country basis. What we should do is look at, for example, the AI Act in Europe, and say, you know, can we take significant aspects of that legislation and put it into place in the US, because all of these companies are global and people are moving around the globe, and companies are trying to reach customers all around the globe. It’s clearly not a state-by-state issue, and is really not even a country-by-country issue anymore. It’s a global issue, and it should be treated globally.
– So then how should people be sort of viewing the potential outcomes of this testimony then, given, as you say, this is a borderless issue and you have different regions of the world approaching this differently?
TOM DAVENPORT: You know, I– my guess is that many of the congressmen and women have bigger fish to fry right now, like, you know, honoring our debt commitments, for example. I don’t expect to see anything happen anytime soon out of this. It would be foolish to ban generative AI tools, because obviously we want to at least keep pace with other countries around the world, specifically China. And they’re doing a lot in this space. A little less than we are, which is one of the few areas where we’re clearly ahead right now. So my guess is that this is more for appearances than for any legislation that might come out of it.
– And when you look at the tenets of the EU’s Artificial Intelligence Act, I mean, there are certain things that they will likely ban, things like social scoring, facial recognition in public, potentially, depending on how it’s being used. What do you think at least would have the greatest likelihood of perhaps getting copied and pasted to the US model?
TOM DAVENPORT: Well, I think both of those things are good. Social scoring, we don’t want to have a social credit score like they do in China, where every time you’re seen at a demonstration or whatever, your social credit score goes down. I think there will probably be an adoption at some point of legislation related to the use of these tools in hiring, as the New York legislation already has. If you’re going to hire someone or reject someone because of an AI recommendation, you have a number of restrictions on how you can do that and how you talk about it with the potential job candidate. So those would be two areas that would be useful.
