In a bold move, a left-leaning think tank has proposed a revolutionary idea: AI-generated news should come with 'nutrition' labels, and tech giants should pay up for using content from publishers. This proposal, from the Institute for Public Policy Research (IPPR), addresses the growing concern over AI's role as a news source.
The IPPR argues that AI companies are quickly becoming the new gatekeepers of the internet, and intervention is necessary to create a balanced and healthy AI news environment. They suggest standardized labels for AI-generated news, revealing the sources used, such as peer-reviewed studies and articles from reputable news outlets. Additionally, they advocate for a licensing regime in the UK, allowing publishers to negotiate with tech companies regarding the use of their content in AI news.
Roa Powell, a senior research fellow at IPPR and co-author of the report, emphasizes the need for fair compensation and clear rules to protect the plurality, trust, and long-term viability of independent journalism.
But here's where it gets controversial... The IPPR suggests starting this licensing process with the UK's competition regulator, utilizing its new powers over Google. This could potentially give web publishers and news organizations control over whether Google can scrape their content for overviews. The IPPR believes collective licensing deals would ensure a diverse range of publishers are included.
According to the Reuters Institute for the Study of Journalism, Google's AI overviews now reach an impressive 2 billion users monthly, and approximately a quarter of people rely on AI for information.
The IPPR proposes that copyright law should remain unchanged to foster a licensing market. They also encourage the government to support new business models for news that aren't reliant on the tech sector, including backing the BBC and local news providers.
The report states, "With the right policies, the government can guide this market so UK news organizations can adapt their business models for the AI age, and AI companies can enhance their product reliability by drawing on trusted sources."
To test their theory, the IPPR evaluated four AI tools: ChatGPT, Google AI overviews, Google Gemini, and Perplexity. They entered 100 news-related queries into these platforms and analyzed over 2,500 links generated by the AI responses.
Interestingly, ChatGPT and Gemini did not cite the BBC, which has blocked the bots used to assemble answers. However, overviews and Perplexity utilized BBC content, despite the broadcaster's objections.
The IPPR found that the Telegraph, GB News, the Sun, and the Daily Mail were cited in less than 4% of ChatGPT's answers, while the Guardian, with a licensing deal with OpenAI (ChatGPT's parent company), was used as a source in nearly 60% of responses. The Financial Times, also with a licensing deal, was featured prominently. The Guardian was the most common source for Gemini, appearing in half of all answers.
Google's AI summaries at the top of search results have impacted click-through traffic and publisher revenues, as many users read the overview without proceeding to the original journalism.
The IPPR raises important questions about how financial relationships between AI companies and news providers influence answers.
"If licensed publications appear more prominently in AI answers, there's a risk of excluding smaller and local news providers, who are less likely to secure AI deals," the report warns.
While licensing deals could partially replace lost advertising revenues, the IPPR believes they won't sustain a healthy news ecosystem. They argue that news organizations could become dependent on tech giants for revenue, and that income could disappear if copyright protections are weakened.
The IPPR proposes public funding to develop new business models for investigative and local news, whose sustainability may be at risk due to the rise of AI news. They also suggest funding for the BBC to "innovate with AI."
This proposal sparks an intriguing debate: Should AI-generated news be regulated, and if so, how? What are your thoughts on the matter? Feel free to share your opinions in the comments!
