Why AI’s Marketing Applications Are Narrow
But the technology's role is poised to expand
Artificial intelligence (AI) is a hot topic in the marketing industry even though many marketers feel unprepared for it. Nikos Acuna, chief visionary at Sizmek, spoke with eMarketer’s Ross Benes about how marketers may use the technology in the future.
How are marketers using AI today?
AI is being used in recommendation engines. Marketers are also using it within predictive analytics. And there are several companies aiming to use it within a real-time bidding infrastructure.
Can you explain?
I’m talking about applying a specific score to an impression opportunity. Right now, most companies are dealing with broad buckets of inventory. So they want to use hard constraints on a specific campaign. They want to target people in specific geographies or specific types of locations.
That’s very different from an AI-driven approach. A true AI approach would ultimately analyze [where marketers can most efficiently drive conversions] and then on its own, adapt in real time to achieve those outcomes.
Could AI be used to help marketers verify whether they’re bidding into first-price or second-price auctions?
An AI that you can teach to play a game of chess and get checkmate every time can’t buy inventory at scale within a real-time bidding environment.
That’s a big topic right now as marketers consolidate the number of demand-side platforms that they are using. The rate at which marketers use AI to determine auction type or to figure out header bidding, the jury is still out on those specific issues. But it can definitely be applied if it’s set up properly to understand the broadest possible sense of data sets.
Where do you think AI will face limitations?
AI is limited in everything that it does. It's inherently narrow.
An AI that you can teach to play a game of chess and get checkmate every time can’t buy inventory at scale within a real-time bidding environment, nor can it learn how to talk to you the way that Siri does. There are a group of scientists that are working on making AI more versatile, but I would say AI is getting stronger within its narrow capacities. It’s absolutely going to get better the more data you push through it.
Is AI the emerging technology that you’re most bullish about?
Oh, definitely. Blockchain is on my periphery in terms of its ability to use network integrity, but that's going to take awhile and it's going to be even more fragmented than the utilization of AI.
What emerging technology do you think is most overhyped?
Everything is actually overhyped. But for those same people that call something overhyped, I challenge those folks to contribute to the discussion in ways that are going to deconstruct what is being overhyped. So I'm kind of the wrong person to ask that. I still think that AI is overhyped, and blockchain is as well.
You think AI is overhyped even though you’ve just been telling me how great it is?
Any new technology is going to be a double-edged sword. This goes back to the discovery of fire. The same thing that helped us cook food was also used to burn villages.
Exactly.
Aren’t you part of the reason then why AI is overhyped?
Probably. I definitely am bullish about AI, yet at the same time when people call it overhyped, I'm not the one that's going to call it overhyped when everything is overhyped.
You seem very enthusiastic about new technology. But tech utopians are eating a lot of crow right now since we’re seeing social media empower white supremacists and large tech companies getting pulled into congressional hearings for failing to police themselves. Is it time to tamper down the exuberance over new gadgets?
If you think about it, any new technology is going to be a double-edged sword. This goes back to the discovery of fire. The same thing that helped us cook food was also used to burn villages.
We need to be more responsible. We need to be more cognizant and find early triggers before they have detrimental consequences to society.