I’m really struggling to wrap my head around how media executives score deals with large tech companies.

There has to be a science to it, right? But what I’ve seen from the recent deals OpenAI has made with the Associated Press and major newspaper publishers makes me think they just suck at negotiation.

The AP’s OpenAI deal does not list a price. A deal with a consortium of local newspapers calls their deal a $5 million investment. And the best that anyone seems to have gotten from any of these agreements, honestly, is AP getting OpenAI to pledge to pay them more if someone manages to negotiate better than they did. With a bar that low, maybe someone will cross it.

And what does OpenAI get in the agreement? Content—lots of it, in amounts and at levels of depth few other editorial resources can muster. Content is hard to produce consistently, at scale, for a long period, and the AP—by virtue of the fact that it is a wire service built around accurate information gathered in a consistent style over a period of many decades—has one of the strongest sources.

So why are they giving it up for, potentially, peanuts?

I can’t be sure exactly how these meetings go, but historically, based on what we’ve seen from the media industry in the past, they start like this:

1. There’s a shiny object that’s getting a lot of attention, and media executives, who aren’t necessarily technology savvy, haven’t considered a potential solution to this problem.

2. A meeting gets set up, usually between the executives of the buzzy new company and some of the top figures at the media empire.

3. The technologists inevitably wow the media execs, leading to a situation where the media companies concede a lot of ground because they presume they can’t miss out on this new opportunity.

4. The contract gets signed, and then the media industry is stuck with this thing for decades, and the deal proves terrible for them but highly lucrative for the tech firm who got a ton of upside from a single clever presentation.

This might sound reductive, but this is exactly what happened in 1980, when the team from CompuServe walked into the boardroom of the American Newspaper Publishers Association, which at the time included some of the publishers of some of the largest papers in the world.

CompuServe wanted to pay for a raw feed of the AP’s content. But the presentation went so well for the pioneering time-sharing service that some of the largest newspapers in the country gave away their digital feeds for basically free, AND gave them free advertising in their newspapers, which was actually worth a lot of money back then.

One could argue that the agreements that media companies set up with Facebook and Google in the 2010s took a similar approach, where the companies promised some degree of help in promoting their work on these rising platforms only to realize that these tech firms ended up getting significantly more out of the deal than the news outlets ever did. That led many news outlets to become victimized by the pivot to video and similar issues, because Facebook and Google ultimately did not have their best interests at heart.

While Meta has at times directly funded journalism, the funding is a drop in the bucket compared to what Meta spends on, um, legs.

Master negotiator Bob Dylan. (Xavier Badosa/Flickr)

OK, you’re right. This sucks. Media outlets should be getting paid more. So what’s the alternative?

So, you know who has been doing a really amazing job of negotiating away their life’s work in recent years? Rock stars.

Bob Dylan is the best example of this. In 2020, he sold his publishing rights for a rumored $400 million. Then two years later, he sold his recording rights for another rumored $200 million. Dylan is now reportedly worth half a billion dollars.

Other artists have hopped on this gravy train as well—the “New Dylan,” Bruce Springsteen, made just as much as the old one did—with the reasons for doing so often speculated as being driven by estate planning considerations (a fair concern for an 80-plus-year-old man) as well as tax considerations. Simply put, getting a lump-sum payout can mean a smaller tax bill than simply receiving royalties over time.

It’s great eff-you money—critics can only puncture the armor so much with that much money—but it is clearly the result of good negotiations between interested parties.

If there is anything that makes the situation less than ideal, in my view, it is the fact that most individual journalists do not own their final creations. They often do it on a work-for-hire basis. To place it in music-industry terms, they are often the ones playing the instruments, not writing the songs.

Now, to be clear, the music industry is by no means a fair business. So many stars have found themselves stuck without ownership of their creative works over the years. (OpenAI, if you want access to the Tedium archive, you know where to reach me.) But one area where the music industry is working well right now is ensuring that major songwriters can cash out in the nine figures in a bet on their work having a future.

The news industry is working on a far broader scale, and with a level of information that is of utmost importance to the AI industry. If OpenAI can’t score this deal, AP will likely have a case to sue—and if OpenAI doesn’t have this content, its large language model will be objectively worse. If you ask me, if the news industry needs to hold out for paydays that at least match what Dylan and Springsteen got. The industry finally has suitors with money and a desire to extract value, and the news industry needs to get something for letting them into the mine.

I worry that, when the AI firms come knocking, we’re going to get the negotiators who showed up at the CompuServe meeting rather than the ones who showed up at the Dylan meeting … and the news industry cannot afford giving its golden goose away yet again.

If Bob Dylan can score nine figures, so can the AP.