The Case for AI Guardrails Without a Slowdown

0:00 AI is helping companies build faster. But as competition at the frontier intensifies, should there be limits on how much risk companies can take to stay ahead? Karen McCormick, chief investment officer at Berengia, argues the answer isn't slowing innovation, but establishing common guardrails around testing, cybersecurity, and incident reporting. She joins us now. Explain that argument. Sure. And, I just prefaced this by saying, so, Beringia, we are early growth investors. So our portfolio companies range in revenue anywhere from 1 to 300,000,000. So they'll be at the smaller end of the spectrum. So we're not gonna be investors in companies like Anthropic or OpenAI, but our portfolio will be on the receiving end of what they're publishing. So I guess our perspective comes from from, those types of businesses. And I would say the most important things to our companies are things like what's the enterprise capability of what's being released, what's the speed at which they're operating, what are the new features that we're seeing, and really importantly, cost. Bundle all of that with, is this product safe to use? Because for companies like ours, we don't necessarily have large teams of either security or cybersecurity or even IT. So we're really reliant on what they're releasing for our portfolio companies Right.

1:13 Who have been net beneficiaries of all of it, but still concerned. So that that's why I think we should talk about regulatory capture. Right? Earlier this week, I was talking to David Sachs about why there should or should not be a greater degree of government interventional regulation, and his viewpoint is right. The frontier labs, they know what they're doing. They should take on the responsibility of safety themselves. But that regular capture regulatory capture, how, an idea for the whole industry serves maybe just two, how would that impact your portfolio?

1:44 Yeah. It's a good question. It was a great interview, and I don't often agree with David, but I did in this particular instance. It does feel like we have, especially the two front runners, Anthropic and OpenAI, who are at on the one hand saying, we really need more safety guardrails, and on the other hand, still continuing to produce at pace. So, I do agree with him that there needs to be some sort of self regulation, and our companies need to be able to trust that the products that they're using are going to be safe. Not the least because due diligence, when people are looking at investing in our companies or acquiring them, they need to have some kind of, safety due diligence and and guardrails in place themselves. So I I think I overall agree with the sentiment that there can be some self policing in the industry. However, one of the bigger risk is when things go wrong, what we tend to find is that government gets more involved. And government regulation for our portfolio companies often means more cost, slower dispatches, and just higher cost, less competition.

2:41 We're talking about early stage investing in AI in private markets. Right? Is there any risk that the environment we're in results in in your section of of the venture market slowing down? Just stop writing checks for a while. I I mean, I would say we already are seeing some of that kind of venture capitalists and and private equity sitting on our hands. I mean, started with SaaSpocalypse and, you know, are all these companies gonna be, decreasing in value because AI can do all of the things that they were doing. We're finding that that didn't really happen. There wasn't quite the SaaSpocalypse that we talked about, but there's definitely a lot of delayed decisions waiting to see what happens. You know, is Anthropic's next release of whatever it is, Anthropic for financial services gonna take the place of a lot of software companies. So I think that's already happened. I don't think it's really a result of guardrails or lack of guardrails, but we are seeing, for example, a lot of our portfolio companies now starting to question and realize what the actual cost of this usage is. And the honest answer is we don't know.

3:37 We know that it's going up consistently, but we don't know what kind of there is no endgame, so we don't know where it's gonna be headed. There's a really big focus right now on the absolute frontier. Right? That's just Yeah. Fact, the most capable. But what about AI at the other end? What about AI for everyday people who don't need it to be spectacular? Is there an opportunity in that? Absolutely. There is a 100%, and I think we focus a lot on, you know, the front runners and and the most expensive versions which have the best capabilities. But the reality is on a day to day basis, whether it's an end user like you or me or our portfolio companies, there are a lot of use cases that don't necessarily require the most expensive version of models. And to be honest, this is where, for example, things like deep sea come into play. Really low cost models for everyday day to day tasks, but there is a question mark around the security level and whether policy is gonna allow us to use some of those types of models. But a 100%, there are different different capabilities, different models for different use cases, and that does help keep the cost down for sure.

4:40 Karen, you you're coming to us from Boston, but you're really focused on The UK and Europe. So explain that part of it. But everything we've discussed, how is that playing out in that market different to The United States? Yeah. It's a really good question. So as you mentioned, Beringia is both US and, European investors. So half of our portfolio is in The US, half is in Europe. We are definitely finding that Europe, which is not unusual, is slightly ahead of the curve from a regulatory and safety perspective. There's more legislation that has already happened, and there's a bit more forefront of asking about the AI safety regulations. I'm a cofounder of a business called Reframe Ventures, which is effectively looking at both ESG and safety requirements for AI into early stage companies. And the Reframe Ventures framework has been adopted widely within Europe, and I would say there's just a more, I guess, forward view of what are we looking at, what are the risks and controls that we need to be looking for within the European ecosystem, although it's absolutely increasing in The US. And we are finding that in The US, whether it's limited partners looking at us for investment or, investors looking at our portfolio for investment, there are a lot more questions now about AI safety and regulation within the companies.