Publishers Will Fight For Compensation From AI, But It Won’t Matter
The Wall Street Journal said it best with its headline about publishers preparing for a showdown with various platforms over their AI tools. And while I am becoming very fascinated by the use of these tools to streamline operations, I agree that publishers should act.
But first… A note about our sponsor, The Content Technologist.
Understanding Google Analytics 4 is a short, daily email course designed specifically for audience editors, publishers, content marketers, and anyone else who has only thought of themselves as a “words person.” Transition to the new GA4, and understand the value of your website’s content.
Pass the GA4 certification and understand digital performance measurement for publishing and media operations.
Presented by The Content Technologist, a resource for content professionals working in the age of algorithms.
Now let’s jump in…
According to WSJ:
In recent weeks, publishing executives have begun examining the extent to which their content has been used to “train” AI tools such as ChatGPT, how they should be compensated and what their legal options are, according to people familiar with meetings organized by the News Media Alliance, a publishing trade group.
“We have valuable content that’s being used constantly to generate revenue for others off the backs of investments that we make, that requires real human work, and that has to be compensated,” said Danielle Coffey, executive vice president and general counsel of the News Media Alliance.
News Media Alliance’s take is right. As a content creator, we should be able to reap some sort of reward for our investments. But what these AI tools do is learn from the content and then regurgitate it for readers. This cuts the content creator out of the conversation and any monetizable opportunities.
This is different than the association’s stance that platforms should pay for linking to our sites. Since launching AMO, I have argued that this is an entitled position to take. We would never promote someone’s content on our sites without being paid, and yet, that’s what happens with Google and Facebook. They don’t charge us to distribute our content. On the contrary, they send us a ton of traffic that we can then convert and monetize.
I take this stance because everyone benefits. The publisher benefits from free traffic; Google benefits from being a great resource to its users. It’s mutualistic.
But the issue with these AI tools is that they are not nearly as mutualistic. On the contrary, there are parasitic elements to them. AI comes, learns, and then leaves, providing the knowledge to readers without needing to direct people to the publications. Even if it provides sources, the intent of these chat-based platforms is to not have the user leave.
So, what can publishers do? According to WSJ:
Reddit has had talks with Microsoft about the use of its content in AI training, people familiar with the discussions said. A Reddit spokesman declined to comment.
Robert Thomson, chief executive of The Wall Street Journal parent News Corp, said at a recent investor conference that he has “started discussions with a certain party who shall remain nameless.”
“Clearly, they are using proprietary content—there should be, obviously, some compensation for that,” Mr. Thomson said.
The publishers will attempt to charge licensing fees. However, how will they determine what to charge? I asked a friend of mine who has experience licensing content how he would come up with pricing. He said:
Another way to think about it is to consider how advertising supports a licensee’s model. So for example, if you charge $2,000 per month and allow 3 articles per weekday, that’s like $33 per article. At a $5 ecpm, will 6,000 people see the ad associated with that article?
So, does NewsCorp do a deal with OpenAI (creator of ChatGPT) where it gets a cut of every search? Let’s look at some numbers.
According to OpenAI’s website, a 24,000 word prompt and response would cost $0.12 for API usage. Assume that your typical prompt and response is 750 words, you’re looking at 32 questions/answers. That’s $0.004 per question.
But let’s keep digging. According to HubSpot, “it’s estimated Google processes approximately 63,000 search queries every second, translating to 5.6 billion searches per day and approximately 2 trillion global searches per year.”
5.6 billion searches at $0.004 per prompt/answer is $21 million per day. Over a year, that’s approximately $7.6 billion. I recognize it’s not a perfect comparison since we’re talking about costs for API usage. But how much should publishers get of that $7.6 billion?
The bigger question is which publishers should be paid. NewsCorp is much bigger, so it should get a bigger cut. But what about niche players? How will each and every publisher get their cut for contributing to the AI’s knowledge? When ChatGPT finally writes the perfect response about why publishers should be capturing first-party data, is A Media Operator going to get a cut? I don’t have the resources to fight for that money.
Sam Lessin, General Partner at Slow Ventures, has what he calls Robots.txt 2.0. The full thing is worth reading on The Information’s forum, but this part stood out to me:
The future for Robots.txt / machine access of websites should / needs to probably look like this:
(A) To access a website as a machine you explicitly register with the service just like a human would. You list who you are, your intended use for data, and obviously what IP addresses you will be coming from / how to validate it is you.
(B) Machines have basic default ‘terms’ that they explicitly sign in order to access a website in terms of how they will access it, how they are allowed to use or retain data, TTL, etc. and maybe even economic terms associated with basic access — all in machine readable format of course
(C) Specific services / companies might negotiate one-off deals for different types of access and use-cases … this will be hard to manage by hand, but you can imagine platforms of ‘smart agents’ negotiating these deals automatically / doing economic and term hand-shakes.
In some respects, what Lessin is describing here is a programmatic-like solution. Rather than me doing a deal with every single AI platform, I could sign up for a third-party platform and let it work out the details. But there are two truths about programmatic anything. First, the middleman takes a solid cut. Second, you never get as much money as you think.
We’re moving into a new era for media. I don’t say that lightly. But the game is changing. If AI-like platforms disintermediate publishers, we may make money from these programmatic solutions, but it won’t be anywhere near as valuable as the traffic we once got. Therefore, we are going to need to develop new strategies to acquire readers. Brand building is going to become far more important.
This could work in niche operator’s favor, though. By focusing on specific audiences, we can go much deeper on certain topics than any other publisher can. Over time, we’ll develop brands with our target audience that will have them coming to us directly. This won’t be easy. For newer operators, it’ll take longer to accomplish our goals.
And it’s not all doom and gloom. None of this piece talks about how, as we build our brands, publishers could introduce their own chat-like tools to provide information to subscribers. If the future really is chat, we should be ready.
Thanks for reading today’s AMO. Become an AMO Pro member to start receiving the premium newsletter on Fridays and for an invite to the AMO Slack. Have a great week!