One by one, artificial intelligence is reviving the debate that the advent of the Internet put on the table. With some schools banning it for fear that students will use it to cheat, and the music industry’s first claims that it can create songs without their control, it’s press time. The Association of tion Media (AMI), an employers’ association that brings together Spanish print titles, has called on Google and Microsoft to sit down and negotiate how they plan to pay for the use of AI in their news.
“AI presents new opportunities, but there is a real risk that they use content without recognizing and rewarding it,” said Irene Lanzaco, director general of AMI (Association of tion Media), in a statement at the sector forum gathered on the portal. A network of journalists. “For this reason, I encourage the major technology companies to start conversations with AMI about the right to use the news from the news media,” he added.
AMI is the new name of AEDE (Spanish Association of Newspaper Editors), which gave its name to the famous one. Canon AEDE Established by Mariano Rajoy in 2014. It was an attempt to charge Google News to link to their news, although the recipients would not be the media itself, but CEDRO. Press SGAE which would then distribute the benefits among the titles. The AEDE law failed: Google refused this payment system and closed news in Spain. in 2022 Congress struck down the AEDE Act and allowed Google to negotiate with each division without Sedro’s mediation, leading to the return of Google News after eight years.
Intellectual property is one of the controversies opened up by artificial intelligence that can generate text. These systems reproduce human handwriting based on a statistical analysis of which word is most likely to come next. Also from media articles.
The new AMI request comes at a time when Google and Microsoft are adding generative AI tools to their search engines, but when they are not yet fully operational in Spain. The most advanced is the company founded by Bill Gates, which added an improved ChatGPT model to its Bing search engine. However, the system suffers from regular service interruptions and has not been able to boost Bing usage globally: respectively statisticianIts market share in March was 8.2%, compared to 8% registered in March 2022.
Bard, Google’s answer to Microsoft’s move, is in a closed testing phase, with access pending. The tool isn’t available in Spanish, and the company hasn’t made any headway when it might be.
elDiario.es contacted both multinational companies to find out about the AMI request. “We have a long history of developing tools to help news publishers monetize their content and strengthen relationships with their audiences,” responded a Google spokesperson, who expects the company to continue to “provide a prioritized experience that sends valuable traffic to the news ecosystem.” It develops new AI tools.
“Very soon, and we’re continuing to work with the ecosystem, including news publishers, to get their input,” a Google spokesperson concluded. Microsoft did not respond to a request for this information as of the closing date.
AMI’s position is not shared by other media associations. The Club for Open Publishers (CLABE), which brings together about 1,000 titles from 180 publishing companies, most of which are digital natives, shares Google’s position. “We don’t want to get into this fight yet. The question of how AI content is created is a debate for all citizens, not just the media,” CLABE Director General Juan Zafra explains to this publication.
AI training, melons to open
AI’s use of copyrighted content isn’t just in the written field. The first to open the debate were several illustrators who saw how generative systems could create new images based on their work. The dilemma works both ways, as it is also unclear who owns the copyright (or responsibility) for AI-generated content.
The details of this training are at the center of most conversations. Typically, the developers of this technology do not disclose what content they feed their artificial intelligence to learn how people express themselves. However, it has been known to be a rough exercise, with abusive, racist or macho texts thrown in. Companies have put filters in place to stop this part of AI learning from being exposed, but these measures don’t always work, so these systems can spread hate and misinformation.
To all this we have to add a new attack launched by European privacy regulators. They launched a large-scale investigation against ChatGPT to find out whether their systems used personal data of EU citizens in training and whether they had permission to do so. The Italian regulator has decided to ban ChatGPT in the country pending the completion of this investigation, after finding signs that this AI is violating privacy laws.
Source: El Diario