Recent breakthroughs in artificial intelligence have raised fresh questions over the impact the technology could have on content production for publishers (and such as BuzzFeed and CNET). In turn, this has triggered debate among marketers as to how the influx of this content impacts ad-buying decisions.

Unlike journalists, marketers have so far expressed little anxiety about the advent of AI-generated content, according to interviews with industry executives. For the advertising community, the means of production pale in comparison to the nature of the content itself. 

“If you are buying TV inventory, for example, you don’t buy against the experience of the showrunner or the time it took to write the script,” said Kate Scott-Dawkins, the global director of business intelligence at GroupM. “You buy based on engagement, ratings and reach.”

The emerging debate over content production mirrors long-standing concerns in the industry about the scope of automation, which has reshaped digital media over the last two decades, said Ana Milicevic, the co-founder and principal of Sparrow Advisers.

Like programmatic advertising, which overhauled how ads are bought and sold on the open web, AI represents the latest in a series of innovations whose transformative capacity is matched only by the uncertainty surrounding its potential.

Labeling AI

Whether a human or a robot created a piece of content matters less to marketers than whether it has found an audience of captive readers, said Ray Rosti, the chief digital officer at Publicis Health Media.

If a piece of AI-generated content resonates with readers and passes standard verification and viewability protocols, most marketers will transact against it, added Scott-Dawkins.

This means that readers play the primary role in determining whether publishers adopt or eschew AI-generated content, which has led to a rise in calls for such content to be labeled or bear a watermark.

Do you know what’s already on the internet? Hundreds of thousands of websites of plagiarized content.

Nandini Jammi, co-founder of the industry watchdog Check My Ads

Marketers would also benefit from knowing whether AI generated the inventory they are bidding on, Rosti said, especially as they look to measure and compare the reach and engagement of such content.

“In the spirit of transparency, it would be ideal to understand when AI is used to create content,” he said.

CPMs, fraud and brand safety

The use of AI to generate content will vastly decrease production costs, making it simpler to scale, according to Rosti. The corresponding uptick in available inventory could, theoretically, deflate the price of CPMs.

But the engagement a piece of content drives, rather than its mere existence, determines its value, said Milicevic, meaning an abundance of cheaply produced inventory is unlikely, on its own, to affect the market.

Worries that AI-generated content would exacerbate digital fraud are also misguided, said Nandini Jammi, co-founder of the industry watchdog Check My Ads.

The internet has no shortage of mediocre content at scale.

Jon Roberts, the chief innovation officer at Dotdash Meredith

Broken programmatic pipelines are responsible for misspent ad budgets, not the ability to produce cheap copy.

“Do you know what’s already on the internet? Hundreds of thousands of websites of plagiarized content,” Jammi said. “AI is just a faster way to copy and paste.”

Likewise, brand suitability considerations depend less on the creator of the content than on the opacity of the programmatic ecosystem. Even as brand safety vendors work to better navigate the challenges posed by AI, humans and robots will never fully rid their work of errors.

As a result, said Jack Smith, the chief product officer at DoubleVerify, their primary focus remains to ensure that the correct ads are delivered to their intended destination—copy errors or not.

That said, reports of AI creating or spreading misinformation at a great pace mean marketers and their vendors need to keep vigilant about content adjacency.

Even as AI evolves beyond its current scope of financial reports and sporting event summaries, the technology will remain unable to perform basic tasks, such as testing a recipe or sharing a personal memory, said Jon Roberts, the chief innovation officer at Dotdash Meredith.

As a result, AI might be able to flood the market with low-quality inventory, but it will pose little threat to premium publishers, according to Roberts.

“The internet has no shortage of mediocre content at scale,” Roberts said. “This is a new version of an age-old temptation.”