Login is restricted to DCN Publisher Members. If you are a DCN Member and don't have an account, register here.

Digital Content Next


InContext / An inside look at the business of digital content

Friend or foe: is AI the latest big tech assault on media revenues?

May 8, 2023 | By Chris Hogg, Chief Revenue Officer – Lotame @Lotame

Every once in a while, a trend rippling through the digital media industry that springs up from the trade publications and into general-interest media. The launch of ChatGPT has been one of those times. AI content generation feels like more than ones and zeroes to media consumers at large. It feels tangible, even visceral, because it has direct ramifications on the production of content itself. Clearly media leaders need to make their voices heard in order to help guide the conversation for users as well as AI developers.

It would appear every Big Tech business wants in on AI chat to power search. But like identity and privacy, media companies are not about to sit back and watch which of the tech giants “wins” generative AI. They need to understand how AI will change content consumption, how they can use it to their advantage, and how to keep walled gardens from absorbing more of their revenue.

A new era for the search business

We’re seeing real and atypical disruption today in AI chat’s applications for search. Bing’s ChatGPT search function removes the onus of poring over multiple search results, and simply generates what feels like a coherent response upfront – be it right or wrong, strong or flimsy, fair, or biased. And history shows that human nature tends to prefer convenience over perfection. Search providers are already investigating how AI search may be monetized, to avoid losing out on the serious revenue search advertising traditionally has delivered (58% of Alphabet’s total revenue last year, for those keeping score).

That brings us to media companies’ reliance on search traffic to monetize their own sites and content. The prospect of users being disincentivized from visiting publisher sites to verify and provide context for information looks ominous to publishers. Today, 26% of all publisher traffic comes from search, in an environment where open web publishers are already competing against walled gardens for traffic and advertisers are reducing their budgets. And as it stands, not all AI search tools link back to their sources, and marketers are still trying to understand how the AI search ad experience should appear. And while search bots already crawl publisher pages, AI chatbots present a less balanced exchange between publisher and search engine.

Publishers that control their data have an advantage

Let’s avoid tunnel vision here, though. Publishers’ imperative to soften any blow to revenue from search traffic overlaps with their existing efforts to reduce reliance on revenue from the programmatic open market. Those efforts involve building loyalty among users and advertisers alike. For users, publishers are offering subscription packages, exclusive newsletter content, and the like. For advertisers, they’re offering private marketplaces and direct deals. Publishers can better navigate AI disruption by continuing and evolving those business strategies. They need to position themselves as sources of trustworthy, high-quality content worthy of the lifetime audience loyalty that can grow CPMs and incentivize deeper advertiser relationships. And AI has real benefits for those strategies.

If and when search becomes a less reliable source of traffic for media companies, they’ll need to ramp up efforts in driving traffic from channels such as social and video. In finding the best channels and business partners, they can turn to AI to analyze large volumes of first- and second-party data, and to do lookalike modelling. Leading publishers have already been doing this as part of their data strategy for withstanding third-party cookie deprecation. Advertisers are taking bold data strategy steps, too, and have the resources to be very advanced here. Publishers should compare notes with their brand partners to learn how to best leverage AI and machine learning for predictive audience creation, modelling, data cleaning and processing, and probabilistic matching.

We’re also seeing publishers take interest in generative AI’s capacity to accelerate and personalize content production, with human staff overseeing and completing the job. AI can do some of the lifting on lower-level SEO writing, bolster production of sponsored content, and source background from the publisher’s content archives. With ad revenues dipping on account of economic pressure, publishers are actively seeking new workflow efficiencies and content initiatives simultaneously without cutting headcount.

IP law will establish new limits for AI crawling

While we’ve all heard chatter about AI displacing human creative talent, media companies have more leverage than they might appear from a slight distance. That human-created input that drives AI output is guarded by intellectual property law. Rights holders are pushing back, and we can expect to see more robust IP regulation, and even technology that protects content from being scraped by AI tools, coming into play in the near future.

Publisher trade body The News/Media Alliance intends to pursue Big Tech companies for IP infringement, and its members want generative AI providers to agree to pay to licence their content (there’s a precedent for such licensing terms, from the use of text snippets in traditional search results). It’s questionable whether scraping publisher sites to train AI would ever fall under “fair use.”

But publishers shouldn’t simply wait for regulators to clear up these issues. That’s a path with unpredictable length and outcomes. At the top of publishers’ to-do lists should be bracing for shortfalls in search traffic, powering their data resources, and finding applications for AI to embolden content production. And Big Tech businesses should be forewarned: AI can’t be a higher priority than the sustainability of the digital publishing business that generative AI models depend on. An input of quality content makes the difference between useful output and unappealing nonsense.

Liked this article?

Subscribe to the InContext newsletter to get insights like this delivered to your inbox every week.