Dazzled by the potential of generative artificial intelligence, we often tend to forget one aspect: i billions of data on which the various models were trained. Data used without respect for copyright and for which the companies that create AI models they never paid.
Now though OpenAI Of ChatGPT he would have changed his mind, and would have signed a multi-million dollar contract to train your model on data from the German publisher Axel Springerowner of Business Insider, Politico, Bild, Welt and others (do you know how to create social media posts with ChatGPT?).
In fact, perhaps you will not remember that AI, in order to learn to draw or converse with us, had to “learn” from something, and companies like OpenAI they used the most accessible source, i.e Internet. The problem, however, is that in doing so they used the intellectual property of millions of content creators, people who have not been paid for their involuntary contributions.
In the last year, with the explosion of artificial intelligenceseveral lawsuits have started for this very reason, and even Europe has imposed rules for companies creating AI models.
OpenAI, like others, it tried to address the issue, but always in an oblique way, for example by allowing the content creators not to let their creations be used for the training of AI, which is certainly not a solution.
There is also another problem, the source and the quality of news. In recent months there have been several reports indicating how the quality of content produced by ChatGPT stay deteriorating, and this despite i heavy investments of OpenAI to introduce new features, updates and models.
The trial apparently influenced the user base of the platform, which has decreased for three consecutive months, and if we add to this the rumors that OpenAI is on the verge of bankruptcy, we understand that we are faced with a big problem.
Now however, as reported by Bloombergthe creator of ChatGPT would be negotiating a deal tens of millions of euros with the German publisher Axel Springer.
According to this collaboration, OpenAI will have access to all the contents of the publishing house, from the archive to current articles, in order to train his models Of artificial intelligence.
According to Bloomberg, OpenAI will provide due credit to Publisher regarding the original source of information, while in turn Axel Springer will leverage OpenAI’s technological expertise to improve its products.
It is rather intuitive to assume that following this new training ChatGPT will not only be able to give more precise answers, but also refer to the original source for greater transparency towards users.
And OpenAI is not the only one moving in this direction. Microsoft is also testing new models for compensate publishers which are used as sources for Copilot, with the intent not only of improve your product, but also to drive more traffic to your publications.
At the moment the deal is not yet concluded, but the COO of OpenAI Brad Lightcap he already declares that he is sure that the partnership will allow users to access contents of quality and real-time news, while ensuring new models of revenue for publishers.
Probably the companies they create AI models they are realizing that if publishers were to disappear, their chatbots they would no longer have sources reliable to offer information to users, and ultimately they themselves would lose a large part of appeal that they have conquered in recent months.
