Should AI have a newspaper subscription too?

Published 7 October 2026 6 min read unprompted
Should AI have a newspaper subscription too?

Should AI have a newspaper subscription too?

When we want information, we turn to various sources, such as social media, newspapers and TV channels. Many of these sources rely on subscriptions, advertising or other forms of revenue to keep their heads above water. But these days, they are no longer the only ways to get information. AI can now collect and process information from the internet using crawlers. As a result, users may be less willing to pay for a subscription. So should AI have a subscription to the newspapers it gets its information from?

What is an AI crawler?

Before we try to understand the link between AI and media platforms, we first need to look at what an “AI crawler” actually is. Traditional crawler bots collect information from the internet and store it in databases and indexes, so search engines can find websites, articles or images more easily. AI crawlers also collect online content, but can use it for different purposes: some gather information to train Large Language Models (LLMs), while others retrieve information and cite sources when answering users’ questions. Put simply, traditional crawlers mainly help people find content, while AI crawlers can help AI systems understand, process or generate content.

The distinction is not always that clear, because some crawlers can serve several purposes at once. Even so, their rise is changing the way people access information. Instead of going straight to a website to read an article, users can simply ask AI to summarise the news. They might visit the original article afterwards, but they might not. In the end, they can get much of an article’s value without actually reading it.

How have publishers dealt with AI crawlers?

Publishers can use a specific file called “robots.txt”. It lets them indicate which parts of a website crawlers may or may not visit. Although crawlers usually respect these instructions, robots.txt is not absolute protection: it is more a request than a technical barrier, and some crawlers may choose to ignore it.

But blocking AI crawlers involves a tricky trade-off. If an AI tool cannot retrieve a publisher’s article, it may also be unable to show or cite the original source in its answer. The publisher protects its content, but at the same time may lose visibility and potential readers.

And this is not entirely new for newspapers. They have had to adapt their business model several times already: from selling a printed paper, to offering news online, and then adding extras such as puzzles, cartoons, recipes and videos. Today, many publishers combine these different products in a single subscription, to offer readers more value than just the news.

What new solutions are publishers trying out?

To stay competitive with AI platforms, publishers have started developing other ways to generate revenue.

Pay-per-crawl is a new system that is still very much in development. It allows website owners to set a price for access by certain AI crawlers. Journalistic content is then treated more like a licensed source.

Pay-per-use is a model in which publishers are paid when their content actually contributes to an AI product. The difference is simple: with pay-per-crawl you pay for access, whereas with pay-per-use you pay for actual use. For example, when a user asks an AI model for news, the publisher is paid for the content used in the answer.

Instead of taking legal action, some publishers choose to make agreements with AI companies about the use of their content. Companies such as Associated Press, Axel Springer, Prisa Media and Le Monde have already signed such agreements with AI companies.

However, this solution is much easier for large media companies than for small newspapers. Smaller publishers may therefore choose to organise themselves collectively to negotiate licences.

This solution may be particularly interesting for smaller publishers. Instead of having to negotiate individually with OpenAI, Google or Anthropic, they can form a collective organisation that manages licences and distributes revenue among its members. This approach is still very much in development, but it is based on the principle of collective management already used in the music industry.

Another way publishers are responding is by making their subscriptions more attractive. Now that people increasingly get their information through social media and AI, fewer people visit news websites directly. According to the Reuters Institute, this makes it harder for publishers to turn new visitors into subscribers.

If AI can tell you for free what happened today, general news becomes less valuable. Publishers therefore need to offer things users cannot easily find elsewhere, such as exclusive investigations, local journalism, podcasts, games or events.

The New York Times is an example of this development. In 2025, only 12% of subscribers paid for news alone, compared with 17% the year before. Most subscribers now choose a bundle that includes games and cooking, among other things.

So, should AI have a newspaper subscription?

Perhaps not in the traditional form. But if AI companies use journalism to offer their products, they should pay for it.

AI is becoming a new way for people to access news, often without visiting the original newspaper or website. If an AI can use an article to answer a user’s question while that user never lands on the publisher’s website or takes out a subscription, the publisher risks losing the revenue needed to produce that journalism.

This does not necessarily mean AI companies should take out a traditional subscription with every newspaper. Pay-per-crawl, pay-per-use, licensing agreements and collective licensing can all offer different ways to compensate publishers. The key thing is that the value does not flow in one direction only: if AI companies benefit from journalism, publishers should benefit too.

AI is changing the way we consume news, so publishers will have to keep adapting. But adapting only works if there remains a sustainable way to fund the journalism behind the information AI provides.