The Financial Times announces it is the latest publication to sign an agreement with OpenAI


Monday, April 29, 2024 at 6:54 pm

The Financial Times (FT) has reached a licensing agreement with OpenAI, the developer behind ChatGPT.

If you can’t beat them, join them. That’s what content providers, including news publications, are currently doing as they work to protect their business models from artificial intelligence (AI) attacks by big tech companies like OpenAI.

Today, the Financial Times (FT) signed a licensing agreement with OpenAI, the developer behind ChatGPT, which will allow users of the AI ​​chatbot to access excerpts, quotes and links from FT articles.

While the value and terms of the deal are confidential, OpenAI reportedly provides news organizations up to $5 million per year to license their copyrighted content to train their artificial intelligence models.

“This is an important agreement in many ways,” said John Ridding, chief executive of the Financial Times Group. “It recognizes the value of our award-winning journalism and It will give us an early understanding of how to present content through artificial intelligence.

He added: “We have long been a leader in news media innovation, pioneering subscription models and engagement technologies, and this partnership will help us stay at the forefront of developments in how people access and use information.”

However, the FT is not the first institution to strike such a deal.

In July last year, the Associated Press took the lead in signing an agreement with OpenAI, licensing its content archives dating back to 1985 to the Microsoft-backed technology company for training. The two companies are also studying “potential use cases of generative artificial intelligence in news products and services.”

Then, in December, Politico and Business Insider owner Axel Springer also agreed to work with OpenAI to let ChatGPT summarize all of its content, as long as it provided linking and attribution.

A few months ago, the Reuters boss said the company had struck “multiple” deals with artificial intelligence companies to use its news content to train their models, but he was tight-lipped on the details.

The boss of News Corp., which owns The Times, said it was pursuing some deals with artificial intelligence companies.

And, while Reddit is not a traditional news publisher, it has struck a deal to allow Google to use its media content to train artificial intelligence tools, reportedly worth about $60 a year, according to Reuters.

How did we get here? Decades ago, when news publishers began moving content online, some initially made it available for free. But the emergence of Google search in the late 1990s highlighted the challenges of maintaining a free model, prompting many publishers to adopt paywall systems.

As media analysts at Enders put it in a new report: “Even if publishers can choose not to have their content used to train new models, they currently cannot choose not to contribute to generating search responses without choosing not to include it. All results in search.

While the dominance of Google search still looms over publishers, the latest threat from Silicon Valley is the emergence of artificial intelligence (AI).

ChatGPT and Google’s Bard train their generative AI models by indiscriminately scraping data from the web.

Content creators are outraged that their intellectual property rights have been and may be used in the future without any consent or compensation.

“The first thing I want to say is they need to start paying us,” the editor-in-chief of The Times said last week, according to the Press Gazette. “They stole our content and we want them to pay for it. Once we figure that out, we’ll be even more excited,” he added.

This has led to publishers negotiating with tech companies to try to strike some mutually beneficial deals.

Like the example above, this has worked. But last December, the New York Times filed legal action against OpenAI and its parent company, Microsoft, after licensing negotiations broke down. The New York Times alleged copyright infringement and sought billions of dollars in damages.

The lawsuit accuses the tech giant of exploiting its “significant investment in its journalism” and threatening its ability to provide “credible information, news analysis and commentary.”

The laws around copyright and fair use in the UK are unclear because they were written before the advent of artificial intelligence. Government plans for guidance on copyright and artificial intelligence initially failed and have since stalled.

The Enders analyst continued: “Publishers currently collectively hold many cards, and while their exact value is difficult to estimate, they should not be abandoned in a hurry.

“For most people, the financial benefits of licensing deals amount to a small increase in income, so terms should be carefully considered consistent with the general commercial goal of maintaining high-quality journalism and online content creation.”



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *

Back To Top