We get AI for work™: New Efforts to Ensure a National AI Policy

Transcript

Eric Felsberg

Principal, Long Island

Hello everyone. Welcome to this special report of We Get AI for Work. Eric Felsberg here with my friend and partner, Joe Lazzaroti. 

We usually have a schedule for releasing our podcast episodes. We thought that it would be important to jump on real quick just to alert our audience about a significant development that happened toward the latter part of last week. That is, President Trump, on the 11th of December, signed an executive order (EO) entitled, Ensuring a National Policy Framework for Artificial Intelligence. Essentially, what the EO is aimed at is the patchwork of state legislation around the country dealing with the issue of AI. We've touched on that during this series as well. The administration's concern is that it creates an inconsistent and burdensome compliance obligation. There are employers out there that would probably agree with that sentiment because it is challenging at times to comply with all these different competing laws. This administration is very keen to see that there are no obstacles to AI development and leadership on the global stage.

The EO has a few features that we thought we would touch on. The first is that it establishes essentially what they would be charged with are challenging state AI laws that conflict with this federal policy of creating a minimal compliance burden for both developers and users of AI out there. First, a litigation task force. The administration will likely go down the road of challenging the constitutionality of these state laws and whether we can have state laws regulating AI. Another feature of the EO is conditioning funds on these states so that they don't have AI laws that are inconsistent with these federal objectives. What the Trump Administration is proposing is that there are certain broadband equity access and deployment funds that potentially would be conditioned on that particular state not having any AI laws that this administration deems, again, an obstacle to this policy. The whole EO runs through the whole thing, which is this issue of preemption. The idea that if we have federal guidance on a particular issue, can a state also regulate that or does the federal law control? We're likely to see some new laws coming out of the FCC, the Federal Communications Commission. With those new requirements that preempt state law, again, some of these questions are yet to be answered, but that is certainly a feature of the EO. 

Where does this leave employers? Well, the idea here again, from the Trump Administration, is to create a uniform framework, which again, as I mentioned, for some employers does have some appeal, but right now, employers are left in a very uncertain position. I know Joe, we've been getting questions like, I saw the new EO come out, so does this mean that I don't need to comply with some of these state laws that are out there. Does this relieve me of that? Right now, the advice has to be that, just stamp out and continue to comply with those laws that apply to you on a state and local level, because while this EO lays out a plan, that plan has not yet been effectuated. We have to take a wait-and-see approach. For the time being, everything is, for now, status quo. 

Joe, do you agree with that?

Joe Lazzarotti

Principal, Tampa

It's interesting because one, typically, when the federal government seeks to apply preemption, there's a federal law that they're relying on to actually preempt. I don't know that there are particular federal laws that they'd be pointing to at this point. The other issue is there's a 90-day period that the EO calls to identify those state statutes that are thought to be onerous for organizations with respect to development and deployment of AI technologies. 

While that may happen in due course and maybe a federal law will be passed, in the time being, if there's something to look at, one thing is perhaps how organizations are preparing policies and contracts. For those companies that have international affiliations or are part of a global company that has policies or rules that have been pushed down to all parts of that organization, including in the U.S., they still may have obligations to comply with frameworks that are embedded in those policies and rules that are coming from other sources. For example, for a long time, the FTC took the position that if a company collects personal data on a website and that website has a privacy policy, the policy that's in place at that time would apply to that collection of data. If we're thinking about how we draft policies or contracts, it's typically not said that this framework will apply only if it's enforceable. In other words, if you say we will develop and comply with these standards, it's by implication that we assume those standards are enforceable. There's this question about how we're drafting things and what rules we're applying by policy that may still apply to organizations, regardless of whether the state can enforce it under some preemption doctrine.

I'm just thinking about how all of these companies are approaching policies, agreements, and rules that are pushed onto certain affiliates and subsidiaries that may actually have some teeth just at the organizational level, not necessarily being enforced at the state level. It's just something to look at in terms of what might be binding the organization and how it uses AI.

Felsberg

That's right, because a lot of the laws in the U.S. that have received the most attention over the recent months have been these bias-audit-type legislation or anti-discrimination requirements, insofar as AI is concerned. If you think about certain states, California comes to mind, for example, where they have an anti-discrimination law and recently amended that law to include AI. Generally speaking, you can't discriminate through the use of AI. With this EO, which we now have from the Trump Administration, how would that work? The discrimination law will still exist. Does that mean that the AI provisions of that particular law would fall under the focus of this particular EO? It's interesting to see. 

One of the things we haven't mentioned is that there's likely to be some legal challenge here, at least now, as we see this. A lot of states may try to challenge the law or the EO itself, claiming that the administration is exceeding its authority, and you have state sovereignty issues conditioning federal funding on some of the stuff. Your point demonstrates this as well. There are a lot of issues right now up in the air that it will be months and likely even years before we see some of these issues actually start to settle.

Lazzarotti

I totally agree. Even existing federal law, how does that play into this? Are the rules governing non-discrimination and privacy affected by this? Typically, one federal law doesn't preempt another federal law unless it's expressly addressed by Congress. There's a bit of confusion around how this might apply, but time will tell, and we'll see what happens. You're right, certainly it looks like this is something that might be challenged by certain states, and that'll just take some time to work its way out.

Felsberg 

Joe, thanks very much for jumping on again this episode. We just wanted to get the word out because we do feel that this is a significant issue for all of our listeners. To the listeners, if you have any questions about anything we've discussed today, and of course, as we always say, if you have any ideas for topics you'd like us to cover in the AI space. By all means, please feel free to email us at AI@JacksonLewis.com.  Until then, hope you all have a happy holiday season. Thank you so much.