Hello, and welcome to this week’s edition of the Future in Five Questions. Brendan recently interviewed Chandler Morse, vice president of public policy at Workday, the software giant that broke into the Fortune 500 last week. The D.C.-based lobbyist discussed the structural issues stopping Congress from addressing the geopolitics of tech, how states are leading the charge on tech rules and why large language models without bespoke datasets are little more than “parlor tricks.” This conversation has been edited and condensed for clarity: What’s one underrated big idea? The role of state government in shaping U.S. tech policy. I think that role is here to stay, and I think it's going to continue to expand. There's this sense across the stakeholders engaged in tech policy that’s almost like this asterisk when it comes to state policy — “it can't be that important.” But I think we're seeing an increasing prominence of state lawmakers in setting national tech policy in a way that, frankly, no one else is right now. Congress is continuing their trend of, "let's think about comprehensive privacy for a few years, let's think about AI for a few years," while these technologies are just moving. States aren't even waiting for the ink to dry on the [European Union’s] AI Act. It's not even implemented, it's got a two-year runway, and they are already moving. We saw Colorado already pass legislation this cycle. I'm bullish about the outcome of AI policy at the state level. In the privacy context, what we saw was a couple of [regulatory] models that industry could kind of deal with. I think there's a real shot at seeing one [AI] model among states. What’s a technology you think is overhyped? Internet-trained large language models with unconstrained prompts — the impact of it is just overblown. So much of the conversation in the policy area is on the existential threat of these large language models. And at least where the market’s going, an unrestrained prompt with a large language model trained by traipsing across all of the data on the internet, I think it's eventually just going to be a parlor trick. Maybe freshmen in college are going to use it to do a first draft of their term papers. But in terms of a meaningful impact in people's lives, I just don't think there's a lot of utility to it. The flip side is, taking those large language models and leveraging them on bespoke, known, safe datasets. There's tremendous utility in that. But so much of the focus is on like, "I went to this website and I asked it what stocks to pick." Eventually, that's just not going to be something that is seen in the landscape. What book most shaped your conception of the future? Prediction Machines by Ajay Agrawal, whom I’ve had the privilege of hearing speak several times and interacting with. It was the first book that sort of helped me understand what AI actually is. It's not C-3PO and R2-D2. What AI really does is, it helps organize information and provide inferences based on that information. AI just helps to surface those predictions in a way that can be meaningful and useful, and to help humans make better decisions. Back to the overhype — that takes a lot of the "oh my gosh, what's going to happen with this technology?" out of it, and a lot of the "oh my gosh, look at what this technology can do for us" into it. And so Prediction Machines was formative. It turns out to have been very prescient to have gotten tutored a little bit in what AI was. Because it has quickly become pretty much everything we do, or at least related to it, on the tech side. What could government be doing regarding technology that it isn’t? Congress can really wrap its arms around the fact that technology policy, in this day and age, is a global geopolitical issue — full stop. And we've got some structural issues when it comes to tech policy in the United States, particularly on the congressional side. If you want to engage on Capitol Hill around international developments in the tech policy arena that will have an effect on the U.S. economy, you go to the [Senate] Commerce Committee, or the [House] Energy and Commerce Committee. Those are the committees of jurisdiction that handle tech and economy. But if it's got an international flavor, is it really theirs or not? And if you go to the Foreign Relations Committee, the International Relations Committee in the House and Senate and say, “I’ve got to talk about this international tech policy thing,” they sort of say, “that's a Commerce issue.” Committee structure changes are hard, so I haven't haven't gotten to solutioning this yet. Trying to get a nucleus of members that are kind of engaged in this is probably the first step. And then maybe looking at structurally — whether it's a standing committee or whether it's some sort of working group. But there's a message that needs to be sent from Congress, that currently I think Congress is struggling to wrap its arms around. What surprised you most in the past year? How quickly every policymaker in the world started talking about AI policy. I manage a global team. And it went from Europe focused on this, to Congress in the United States focused on this, to there's a bill in Canada, there's activity in Australia, in Singapore, in Japan. There's the OECD. There's WEF. Even our footprint is global, in terms of these conversations. I think the surprise is that it's happening at the same time as adoption. In the past what we've seen is policy conversations lag that adoption. So the technology gets adopted, and then the policymakers start to think about unintended consequences and things they have to fix. This conversation is happening full-bore, full-throat, in capitals all over the world, at the same time everyone is adopting the technology. We think that makes a ton of sense, because we think the more people can trust this technology, the better the outcomes and the better the benefits and the better the adoption. So we welcome that. But it is a surprise, because it has not been how tech policy has been fashioned in the past.
|