With the risk of copyright infringement increasing with the spread of generative AI that automatically produces elaborate texts, some experts say legislation to protect copyrights is necessary.

The Japan Newspaper Publishers & Editors Association has warned that search engines powered by artificial intelligence are “free-riding” on news content in a statement on Wednesday.

With the risk of copyright infringement increasing with the spread of generative AI that automatically produces elaborate texts, some experts say legislation to protect copyrights is necessary.

Copyright infringement

In response to a search for the number of victims of the Noto Peninsula Earthquake that occurred on New Year’s Day, a Google search engine easily returned text such as, “282 people were killed in the Noto Peninsula Earthquake as of June 19.”

The text is believed to have been generated based on multiple reports from TV stations and major newspapers. The description of the cause of their death was almost identical to the wording used in an article carried in a major newspapers.

In August last year, Google began offering this search-linked, generative AI service called SGE (Search Generative Experience), which is built into its search engine, on a trial basis.

With conventional search services, multiple websites are listed according to the keywords entered, and the user goes to the website to view the information.

In contrast, AI-powered search engines convert and process information from multiple websites related to keywords and display the results as a single text.

The Copyright Law states that search services are permitted to use another party’s copyrighted work without permission provided that the use is minor and only to the extent deemed necessary for the purpose of the search.

Yet, the risk of copyright infringement has increased beyond “minor use” because AI-powered search engines generate long texts based on copyrighted articles.

The association’s statement emphasizes the need to improve the legal system including the revision of the law that suits the era of generative AI.

The association also noted that there are cases where incorrect information is displayed, expressing concern that this could lead to disadvantages by providing misinformation to users.

Foundation of democracy

The statement warned that if such free-riding of articles and other content is not regulated, content production will die out, causing “irreversible damage to the foundation of democracy and the nation’s culture.”

Generative AI search engines allow users to get a general idea of information available from the AI-generated answers, reducing the number of people who think of visiting the original websites.

News organizations receive advertising revenue based on the number of views through each news website.

The news media spend a great deal of effort and money on news gathering and reporting the news. Should “free-riding” be allowed, the chances of providing accurate and diverse news coverage could be reduced.

An executive of a local newspaper expressed a sense of urgency, saying, “Our coypyrighted work is displayed in a very similar way,” and “It is like taking bread out of the mouths of our businesses in terms of business and news activities if the service is used for free.”

“If people no longer have access to websites [of the news media], advertising revenue will decrease, and news organizations and other primary information providers will be at a great disadvantage,” said Prof. Makoto Nagatsuka of Hitotsubashi University, who specializes in copyright law. “The legal system should be reviewed in light of the rapid progress of generative AI.”

Other industries, overseas

In other industries and overseas, efforts are underway to protect the rights to copyrighted works as well.

In August last year, the Japan Magazine Publishers Association, the Japan Photographic Copyright Association, and the Japan Book Publishers Association issued a joint statement with the Japan Newspaper Publishers & Editors Association calling for consideration of measures to protect their rights.

They said that the current Copyright Law is favorable to AI learning, but that the relief measures for copyright holders are inadequate.

In May, the News Media Association, an industry group of about 2,000 newspaper and magazine publishers in the United States, asked the U.S. Department of Justice and the U.S. Federal Trade Commission to halt the expansion of AI search engine services, claiming that some operators were improperly using articles from newspapers and other media.

Asia News Networks (ANN)/The Japan News