EXCLUSIVE: At CES 2024, What Retailers Need to Know About AI

The world's biggest technology trade show opened up this week in Las Vegas, and the letters on the tip of everyone's tongue are A and I. 

Artificial intelligence, and specifically the generative AI boom launched by OpenAI's ChatGPT, has prompted everyone from Walmart to TikTok to make a big splash at this year's CES, which is expected to draw more than 130,000 attendees and 4,000 exhibitors, up at least 15% from 2023's event according to Consumer Electronics Show officials.

[Read more: "AI Isn’t Always the Answer"]

The show schedule this week includes multiple sessions on GenAI, but one session on Jan. 8 did a deep dive into two key areas of the GenAI phenomenon that not too many people are talking about, and are of particular importance to retailers.

"I think this year is going to be the year of data and trust," said Steve Canepa, general manager, Global Industries at IBM. Canepa was a panelist in a session titled "The AI Inflection Point," along with Leslie Shannon, head of trend and innovation scouting at Nokia, and Richard Kerris, VP of GM media and entertainment at NVIDIA Corp. 

Canepa leads all industries for IBM and started IBM's global media and entertainment industry business in 1995. 

"Look, there is no AI without data. You have to have an information architecture in order to supply data into models, in order to train and tune models, and in order to get the outputs that you want. Most firms struggle with getting the right data, having confidence that there's one version of the truth, getting those data into the models, etc."

Shannon agreed, saying that "being able to use your data today really depends on having somebody going to the data scientists and telling them what the good questions are. And, quite often, that just doesn't happen. And so there's this kind of unintentional air gap between the people who know the business and the people who can actually query the data."

But with GenAI, Shannon said, and "particularly a trusted generative AI, this means that now using natural language, the people who actually are at the front end of the business, the ones who are really doing this stuff, they can begin to ask questions of the data and get answers. And this is actually going to be tremendously transformational."

The second area of importance, Canepa said, is GenAI cybersecurity.

"The other issue that I think is going to come to the surface in a very big way this year is the bad guys aren't on vacation," Canepa said. "So just like we did in security as the threat service got bigger and bigger and bigger, and we had to get more and more and more sophisticated to identify it, to predict it, to avoid it, similar things are going to happen this year in the way that generative AI gets applied."

 Canepa described a possible attack in detail.

"Some of these attacks do really interesting things, like they put a hidden piece of data that gets scooped up into a model," Canepa said. "And then along comes a prompt. And that data is just sitting waiting for that prompt to come along. When it does, it activates it. And when it activates it explicitly tries to get the model to do something that it's trained not to do through various creative ways. And so, not to be an alarmist, but there are these kind of ticking hidden catalysts that are sitting there."

So to have confidence, Canepa said, "we're going to have to get very good about understanding what was used, as I mentioned before. Because we're going to have some really unpleasant moments this year that come out of models doing things that no one intended for them to do, but they happen."

Kerris, who was chief technology officer at LucasFilm and has also worked at Apple, said that companies are already starting to develop solutions.

"I think each of these things are opportunities for companies," Kerris said. "As we saw with computer viruses and then companies were born out of that to protect your systems and things like that. We will start to see watermarks and other things happen with the content that's created. We're already seeing companies such as Getty, Adobe, and many, many others who are ensuring that their data is clean and then indemnifying the use cases of it."

To close out the session, Shannon asked Kerris the thing he is looking forward to GenAI making possible in the near future, or far future.

"What always excites me is giving the access to tools and technologies that people that have never had it before. Like what we saw in the music industry. You see music now coming from all parts of the globe, and it's incredible. I believe we're going to have storytelling come from all parts of the globe. And what excites me is what are we going to hear about that we never even dreamed about, that people that have access to these tools will now be able to do?"