“The reality check is happening now”: AI leaders on the state of the market at Collision

As expected, at Collision 2024 this week, artificial intelligence (AI) dominated discussions.

A year and a half after the release of OpenAI’s text generator ChatGPT, BetaKit checked in with industry leaders at the Toronto technology conference on the state of the AI market, where there has been real innovation—not just smoke and mirrors—and what they expect going forward.

“I think there is still a lot of [AI] hype, but the reality check is happening right now,” Thomson Reuters chief product officer David Wong told BetaKit in an interview.

“I think there’s going to be a lot of money lost in AI.”

Wong argued that it is “both one of the best times and one of the hardest times to be building enterprise AI.” He claimed a correction is already starting with enterprises. Compared to 12 months ago, corporate buyers have become more discerning, he said.

Now, Wong said, the conversation for smart enterprises is, “Cool, what does [your tech] do, what’s the [return on investment], and why are you better than the five or six or seven other companies out there that are doing, apparently, the same thing?” He noted that these conditions have made things tougher for startups and companies selling AI products.

In an interview with BetaKit, Matt Wood, vice-president of AI products at cloud giant Amazon Web Services, described generative AI as “the single largest shift in terms of how we’re going to interact with data and information and each other, probably since the advent of the very earliest internet and the very earliest web browsers.”

This belief in the technology’s transformative potential has been reflected in the explosion of AI companies and experiments over the past couple of years, as entrepreneurs and executives have sought to apply AI to solve as many problems as possible. It has also fuelled tremendous hype—even as venture capital (VC) has become harder to come by amid the market downturn, big AI winners have been crowned and funding for, and valuations of, AI startups remains high.

Autonomous trucking startup Waabi just closed a $275-million CAD Series B round, marking the second major Canadian AI financing in recent weeks. Fellow Toronto AI company Cohere, an OpenAI competitor, closed more than $600 million in fresh capital. Meanwhile, Tenstorrent is also reportedly seeking hundreds of millions in funding and fellow Toronto AI chip maker Untether AI is looking to raise a nine-figure financing of its own.

“When you parse it out, I think there’s real progress in the underlying hardware, silicon technology,” Chris Walker, an Intel veteran and CEO of Untether AI, told BetaKit in an interview. “There’s certainly been amazing progress in the models, especially in the generative [AI large-language models], that captures everybody’s imagination of what’s possible.”

Wong expects “the real use cases” of AI to begin to shake out “over the next little while,” comparing the current period to the dot-com boom and bust of the late 1990s and early aughts. “Are you going to find the Amazon or the Pets.com?”

As the market differentiates what offers true value from what is smoke-and-mirrors, experts expect some pain to follow. “I actually think—and I may be in the minority for saying this—I think there’s going to be a lot of money lost in AI,” Spencer McLeod, partner at Chicago VC firm G Squared, warned during a June 18 Collision panel. For his part, McLeod believes that the current market features “a lot of companies that are “pretenders,” “charlatans,” and copycats.

McLeod added, “Just because you say AI, ladies and gentlemen, does not mean you actually have an application to AI. I think there is a real bubble that is transpiring in parts of the AI ecosystem that is dangerous, as people chase momentum—both companies in the buying cycle and investors chasing the hype cycle.”

Speaking on stage alongside McLeod, White Star Capital general partner and fellow VC Christophe Bourque noted that every board of a large company is asking their CEO what they are doing with AI. “The reality is that most enterprises don’t have a good answer,” he said. Per Bourque, this uncertainty has trickled down and slowed software-as-a-service deals, with many budgets stuck until corporate buyers’ AI strategies become clearer.

As to where we currently sit in the AI adoption cycle, according to Cohere co-founder and CEO Aidan Gomez, “2023 was kind of the year of the proof-of-concept.” During his June 18 fireside chat at Collision, Gomez said enterprises saw the potential of this technology when it “blew up” in the consumer space following the launch of ChatGPT, began asking how they could adopt it, and spun up tons of proofs-of-concept.

In the process, Gomez said they encountered two main barriers to AI adoption: data privacy and scalability. “I think those two things have held us back, but those barriers are coming down,” Gomez said. “As those barriers come down, and as enterprises get [comfortable] around the technology and build trust, I think we’re going to see [AI adoption] take off and we’ve already started to see the early signs of that,” he claimed.

So far, Bourque has seen AI making more of a difference for firms on the cost-cutting than the revenue-generating side of the equation, citing the example of buy-now, pay-later company Klarna, which has realized big savings by replacing human customer service agents with chatbots. But most businesses, he said, have yet to realize AI’s full potential.

“I don’t think that many companies have really figured out how to get the full value out of this technology,” he argued, noting that these conditions have benefitted consulting firms, which are “making a killing” advising in the meantime.

For his part, Gomez expressed the belief that AI will augment rather than replace workers. “I’m a big believer in augmentation and not displacement, so I don’t foresee people losing jobs,” he said. “We want to be doing more and we don’t have the people to do it, so we need to make the people we do have much more productive.”

The Cohere CEO acknowledged that certain roles and sectors will be affected more than others by AI, but expressed optimism that the tech will be a net positive for workers globally.

According to Wong, customers of Thomson Reuters, which caters to lawyers, tax, and risk and fraud professionals, have been using AI to make information retrieval easier and produce written work across a broad set of use cases. “But it’s not replacing lawyers and accountants—it’s augmenting their work,” he said.

“It’s not replacing lawyers and accountants—it’s augmenting their work.”

“Generally, where we’ve seen it fall down is where claims are made that [AI] systems are doing that intellectual work,” Wong said. “So, ‘It can make the decision, it can formulate the arguments, it can … provide the defence.’ That’s not realistic right now, and frankly, customers won’t trust it to do those types of things.”

For some time now, AI godfather Geoffrey Hinton has been raising awareness about the potential future dangers of AI and warning that the tech poses an existential threat to humanity. Hinton has argued the development of increasingly advanced AI is inevitable, and “The best we can do is try and stay a bit ahead of the game,” in terms of ensuring it is used as safely as possible.

At his June 19 fireside at Collision, Hinton reiterated his previous concerns and emphasized a potentially more immediate and urgent one: the problems associated with combining AI with surveillance, as well as some other use cases.

For his part, Gomez believes that “It’s going to take us a while to exceed human capabilities uniformly,” and is skeptical of the worst-case scenarios Hinton and others have voiced. “There’s still extraordinary things that will happen; it’s just not a doomsday AI scenario,” Gomez argued.

Feature image courtesy Web Summit.