Feb 20, 2024 - Business

Exclusive: NYT plans to debut new generative AI ad tool later this year

NYT app

Photo: Thomas Trutschel/Photothek via Getty Images

The New York Times is recruiting partners to test new ad-targeting solutions using generative artificial intelligence (GenAI) beginning next quarter, executives told Axios. It plans to make the new tool widely available to marketers in the second half of the year

Why it matters: The Times started working on these efforts long before it filed a landmark lawsuit against OpenAI and Microsoft for copyright infringement, executives said.

  • "This obviously demonstrates that we believe GenAI is an enabler and can be something that is effective for our business when used responsibly," said Joy Robins, global chief advertising officer.

How it works: The new technology, which is being created internally at the Times, will deliver a recommendation for where an ad campaign could perform best based on its message or objectives.

  • The tool can also be used to target niche audiences that were previously unidentifiable based on their interests, aspirations and opinions.
  • For example, a car company looking to position itself as family-friendly might use the tool to align some of its messages with articles about technology and others to pet owners across the Times' site and app.

Zoom in: The Times is currently experimenting with multiple large language models (LLMs), including enterprise models and open-source models, to power the new tool, said Mohit Lohia, senior vice president and head of the Times' digital advertising mission. The tech, he added, would work using any LLM.

  • The Times has one of the largest news audiences globally, with over 10 million paid subscribers and over 100 million registered readers.
  • That scale, combined with "phenomenal content" and "very well organized data," gives the Times' ad department a good baseline to begin experimenting with new types of GenAI products, Lohia added.

Details: The Times is currently recruiting select advertising partners to participate in a closed beta test that will likely begin early in the second quarter.

  • It's hoping to make the new tool available to more advertisers after it gets results from its first round of experiments.
  • Beta partners are being pulled from the Times' Product Council, a group of marketers across various industries that the company leans on to help experiment with new products and technology.
  • Brands themselves can't use the tool directly, but they can work with the Times' advertising team to apply it to their ad campaigns.

Between the lines: It's unclear at this point how the Times plans to price ads that leverage the new tool to target their campaigns, Lohia said, noting that those decisions will likely depend on how the experiments perform.

Be smart: The new tool is designed to make it easier for advertisers to plan and optimize ad campaigns in real time.

  • The Times has already built a set of contextual ad-targeting solutions for marketers based on the first-party data it collects from its users, including a tool that can target ads by sentiment.
  • But those offerings require advertisers to explicitly choose targeting criteria before their campaigns run, making it harder to optimize them instantaneously.

The new tool is also being designed to help advertisers avoid potentially awkward alignments against content that wouldn't fit in with their messaging goals.

  • For example, a textbook brand could align its campaign to run next to education content, but avoid articles related to bullying.

The big picture: More publishers are experimenting with ways generative AI can improve real-time ad targeting amid the broader industry pullback from third-party tracking cookies.

  • The Times has been investing in the technology broadly across all of its products, despite its ongoing legal battle with OpenAI.
  • In December, it hired an editorial director of AI initiatives to experiment with GenAI tools in its newsroom.

The bottom line: "This has been a top priority for the team and we've been working on this since the last three quarters now," said Lohia. "It is something which is taking a disproportionate amount of our resources."

Go deeper