Google has long been the go-to tool for search, so how will AI-powered search change that? Dave Winterlich, presenter of industry podcast Inside Marketing, puts the question to Dan Calladine, head of media futures at dentsu.
Thanks to generative AI, the transformation has already begun. Instead of a search resulting in a choice of links, generative artificial intelligence (GAI) tools such as ChatGPT or Google’s Bard present you with a single answer.
“The problem is that potentially, rather like some people we might know in our own lives, if it doesn’t know the answer, it has a tendency at the moment to make things up,” says Calladine.
GAI has also given rise to copyright questions in relation to the data its large language models are trained on. Under traditional search, when we ask for a recipe, we get to choose what we believe are trusted sources. That’s harder to assess with GAI.
Making products that are more relevant, more efficient and run smoother with less energy consumption
How your KitKat break is helping cocoa-farming families close the living income gap through an innovative programme
Renewable gas will power Ireland’s decarbonised future
All aboard: Ten great things to see and do in Heritage Week
Moreover, says Winterlich, when it comes to image generation, questions around intellectual property arise when images are generated that look similar to, but are not of, celebrities.
It’s a grey area, says Calladine, and one which AI companies will increasingly put guardrails around. However, companies such as Adobe, Getty and Shutterstock, which have trained their AI models on their own pictures, will face no such IP concerns.
Good corporate governance will go a long way towards reducing the risks to agencies and clients. “If you are producing anything which is public-facing, then it’s got to be completely compliant,” says Calladine.
Impact of GAI on advertisers
Advertisers are already familiar with the use of AI tools such as Google Performance Max for media optimisation. But many are still trying to get to grips with its true value. “You’re kind of trusting Google to mark their own homework,” points out Winterlich.
But it’s only the start. Meta’s Advantage+ helps advertisers figure out more quickly which ecommerce campaigns are converting.
![](https://www.irishtimes.com/resizer/v2/AQ5IFZ6LDZCW7A7WERBTSMCAI4.jpg?auth=22607bd1fb86e50d7f42039b7eaa12148826bb6ff293c97987b75582c5c10eb1&width=800&height=1066)
“We’ve been using it quite a lot and it seems to give very good results. You hear good stories about increased ROI, and it’s very automated, but you’re ceding quite a lot of control in terms of exactly how your ads look, and potentially what sort of audience you’re going for,” cautions Calladine.
Great if the AI optimisation works, but if it doesn’t, “you’re not really able to see forensically what went wrong with a campaign, and what the learnings are,” he adds.
The rise in protectionism
Expect platforms to become much more protectionist about content. Up until now, open access was the norm for APIs and the “pipes”, or digital connectors, that linked them. This allowed, for example, the integration of content, such as say, Google Maps with Uber.
“What we’re starting to see now is a lot of these apps saying, ‘Actually, we don’t want you to be able to play with our content, or embed our posts onto your site’,” says Calladine.
Either third parties can’t use their content at all or are going to have to pay for it.
One reason is simply to generate revenue. But increasingly they want users to stay on their own app or website, “so they can serve me adverts and monetise me, but also so they can give me a first party cookie and work out how to target the right sort of advertising to me – and that’s obviously very valuable”, says Calladine.
For platforms planning a chatbot of their own, the data held by companies such as X and Reddit are now incredibly useful.
People who wondered why Elon Musk spent so much buying Twitter now look at that transaction differently, given the value in a GAI world of 15 years’ worth of old tweets. “It just gives you so many examples of how people speak,” explains Calladine.
Doubling down on identity
Increasingly, platforms will take back more control over their relationship with users, predicts Calladine.
Netflix clamping down on the sharing of passwords is a good example.
“Obviously it gives them extra revenue. But the other thing is, they simply learn more about the people who are actually watching. Now that they are going after the same sort of ad budgets as TV stations, they really need to know who the audience is,” says Calladine.
“If they can get people who were using their friends’ accounts onto their own accounts, then they’ve got an e-mail address, and a payment card, and can see quite a lot of things about them. That’s really useful data.”
As a result, expect more focus on verified users and premium accounts. “It’s a way of finding out more about people, and monetising them much more effectively,” he adds.
Ad-loads up
Expect more ads. “In the most recent quarter, Meta said the number of impressions they’d shown to users had gone up 31 per cent, but their number of users had only gone up 7 per cent. So clearly everybody’s getting more ads,” says Calladine.
The same thing is happening from Instagram to TikTok.
One of the reasons people started getting fed up with MySpace was because of the increased ad load, he points out, which in turn helped drive people to Facebook.
Platforms will have to tread carefully. “Obviously people will prefer things in some respects without advertising if they feel the advertising isn’t useful to them. But then you look at Amazon. In the last four quarters it has generated something like US$42 billion (€38.4 billion) in revenue, and people don’t complain about the adverts on Amazon,” says Calladine.
That’s because they are already in shopping mode, and receptive, points out Winterlich.
Calladine agrees. “In retrospect, what Amazon has done is an incredible trick. If you’d said five years ago that in five years’ time, there is going to be US$10 billion worth of ads on this site, and in this app, every quarter, you would have thought, ‘Oh my God there will be pop-ups everywhere, I’ll be dodging full page interstitials’. Yet they seem to have done it incredibly well. So there is a way of adding advertising and making it useful,” he points out.
While we tend to think of streaming as a relatively ad-free experience, “advertising is coming,” says Calladine, pointing to moves in that direction at Disney+ and Amazon Prime Video.
“As an agency, we’re not complaining about more opportunities to potentially target people in more contexts, but the problem is the huge amount of clutter there is going to be,” he cautions.
Advertisers will have to gravitate to premium spaces, or places where there is less clutter. Technology that measures attention is going to be increasingly important. But the real challenge will be encouraging clients – and their impression-focused PR agencies – to move.
Bigger won’t always be better
One fear is that AI will benefit so-called walled gardens, the mega platforms big enough to keep people within their “everything” apps. That presents risks to minority-owned media, particularly given the rise of AI planning.
“We need to be really conscious as marketers that we can’t do everything according to the algorithm, or just let the machine go for the lowest-hanging fruit,” says Calladine.
Relying on algorithms to target a particular advert at the people most likely to respond excludes the potential to find new customers. “Don’t just trust the algorithm to automatically find the right audience for you,” he cautions.
To hear this episode of Inside Marketing visit irishtimes.com