Professional media accuse chatbot developers of illegal use of content

The professional association News Media Alliance, representing almost 2 thousand newspapers and news resources from the United States and Canada, criticized technology companies that create chatbots. In his report News Media Alliance experts point out that developers such as OpenAI and Google use press materials to train chatbots, but do not enter into licensing agreements with publishers or pay them for using the content.

The report’s authors note that developers “illegally copied and used content from news resources, magazines, and digital media to train” chatbots. Moreover, the algorithms were set up in such a way that the chatbot trusted information published in reputable publications more than materials appearing elsewhere on the Internet. According to News Media Alliance CEO Daniel Coffey, this shows that developers “recognize our uniqueness and value, but most of them do not obtain the necessary permissions through licensing agreements or pay for the use of content.” And this attitude towards high-quality, human-created content, according to experts, harms not only publishers, but also undermines trust in the artificial intelligence models themselves.

Some publishers have already begun to take steps to protect their content from such use. Thus, in August at least 535 news companies put a block that prevents their content from being collected and used to train OpenAI’s ChatGPT. Meanwhile, as the News Media Alliance notes, this only protects content from new attempts to extract data from their sites, but does not solve the problem of protecting content that has already been taken by developers. To solve this problem, the report’s authors made a number of recommendations for publishers. These include the involvement of the American Copyright Office in the process of resolving the problem, the development of new content licensing models, the adoption of relevant legislation, etc.

Alena Miklashevskaya