The art of negotiating AI deals, according to Time COO Mark Howard

“[In] the negotiation part, of course, you have a much better position if you have the ability to also threaten litigation,” said Howard. But, “there’s a lot of opportunity by being more aggressive and really working together with these companies, as opposed to viewing them as an existential threat to the publishing world.”

During an on-stage session at Digiday’s summit, Howard shared his philosophy on why his company is choosing to work with AI companies, rather than fight against them, and, most importantly, how to best go about the negotiation process.

“It’s kind of like buying a car. The minute you drive the car off the lot, [you think,] they got a little bit of the best of you, but you’re driving away with something that hopefully you feel really good about,” he said.

The earlier, the better

There is not a singular, common model that all AI companies use for deals with publishers, but being first usually means there is a better chance at getting a favorable deal, Howard said — especially if the AI company hasn’t figured out its business model yet.

“Our belief has been, if we can be in earlier, it’s probably better. The longer that those negotiations go on, I can’t imagine the deal terms ever getting better,” he said, adding that’s especially the case once a company’s business model proves to be fruitful and more publishers start asking for deals.

Take ProRata.ai, which is positioning itself as a curated large language model sourced predominantly by premium publishers and content sources. The pitch to publishers seems to be that it will accurately attribute to their content and offer a revenue share from the money the company earns off of its AI chatbot subscriptions, per a report by Axios. But Howard said that, like many other AI companies, ProRata.ai is still figuring out how it’s going to make money, meaning publishers partnering with the company are committing to a vision, not a proof of concept.

“We took a little bit of a leap of faith, but we believe that they will figure out the best path forward, and we would rather be in early thinking about how we show up in that environment, how our brand is represented, the citations and that’s part of the negotiation for each of those deals,” Howard said.

Partners should help publishers understand AI infiltration

What’s more, AI companies already have publishers’ content, Howard said. And knowing how much of it they have is the strongest way to go into the negotiation process.

That’s why Time’s chief technology officer Bharat Krish and chief information officer Sharon Millz have worked to understand how much AI bot activity is on the publisher’s site, already scraping the publisher’s content, he said.

TollBit is one partner that Time uses to monitor AI bot traffic, which Howard said he hopes will eventually allow publishers to charge incremental fees for content scraping and conduct content licensing at scale. But for now, the data that TollBit reports back to Time is leveraged in a lot of the negotiations that Howard’s team has with AI companies.

“There’s no doubt that there’s going to be millions of these bots out there on the internet. … [But] there’s no way you can negotiate with more than just a handful of these companies,” he said. So when it comes to the deals that publishers should prioritize when first embarking on this journey, Howard said to start with the “bigger players” that are already willing to work with publishers.

Traffic is part of the bargaining, but not a guarantee

While Howard said that attribution and links to Time’s content is always part of the discussion with AI companies in the deals his team considers, it’s not a deliverable that’s negotiated into the contracts yet.

But with that said, after all the issues of hallucinations in the earlier days of the technology being rolled out, AI companies recognize the value of citing credible, premium journalism within their AI-generated results, because it gives the technology more credibility with users.

Content has value but there’s no common currency

Howard said he doesn’t believe that there needs to be a price per word or article model when publishers assign value to their content. That’s because that price would vary greatly from publisher to publisher, depending on the depth of their archives and the quality of journalism behind the content.

And the more premium the content, the more value there is for AI companies to gain credibility, Howard said, which is the strongest tool publishers have in their negotiation toolbox. While he said publishers should go into a negotiation with a monetary value in mind and to not settle for less, he declined to provide an exact figure.

“There’s real value and so how do you negotiate something like that when there’s no currency tied to it? It’s difficult,” Howard said.

As for how the most favored nation (MFN) clause gets factored into deals — which means AI companies would have to treat all trading partners equally — Howard said AI companies don’t want to negotiate that way. But that doesn’t mean they should be allowed to low ball publishers, either.

“The media industry has always been a series of Faustian bargains. We’ve been doing this since the digital media really launched. … We’ve all been Charlie Brown and Lucy with the football, and the big tech companies have pulled it out from under us before,” said Howard. “But these companies are very well funded. … The opportunities out there are tremendous, if we can all capitalize on it right now.”