Think Tank Proposes ‘Nutrition’ Labels for AI-Generated News

According to a left-leaning think tank, AI-generated news should be accompanied by “nutrition” labels, and technology companies must compensate publishers for the content they utilize. This recommendation comes as the usage of AI technology as a source for current affairs grows significantly.

The Institute for Public Policy Research (IPPR) has highlighted that AI firms are quickly becoming the new “gatekeepers” of the internet, warranting immediate action to cultivate a healthy news environment influenced by AI technologies.

The report suggests that all AI-generated news should be marked with standardized labels, informing audiences about the sources of information used to generate these news items. This includes referencing peer-reviewed studies and articles from established news sources. Furthermore, the IPPR advocates for a licensing framework in the UK that enables publishers to negotiate usage rights with tech companies regarding their content in AI-generated news.

“As AI companies profit from journalism and shape public perception, it’s imperative they are mandated to fairly compensate the news they utilize and adhere to clear regulations that uphold diversity, trustworthiness, and the sustainable future of independent journalism,” emphasized Roa Powell, a senior research fellow at IPPR and co-author of the report.

The IPPR proposes that work on licensing could initiate with the UK’s competition authority leveraging its new enforcement powers over Google. Recently, the Competition and Markets Authority suggested allowing web publishers and newsrooms to opt-out of having their content scraped by Google for its AI overviews. According to the IPPR, collective licensing agreements could encompass a diverse range of publishers.

With AI overviews from Google now reaching 2 billion users monthly, and about a quarter of individuals utilizing AI platforms for information, as reported by the Reuters Institute for the Study of Journalism, this intervention seems increasingly essential.

The IPPR believes that existing copyright legislation should remain intact to foster the growth of a licensing marketplace. Concurrently, the government should explore innovative business models for news that do not rely solely on the technology sector, emphasizing support for local news organizations and the BBC.

“With appropriate policies in place, the government can shape this market to assist UK news organizations in transitioning their business strategies for the AI era. Simultaneously, AI companies can enhance the reliability of their offerings by utilizing reputable sources,” the report stated.

To assess the current landscape, the IPPR evaluated four AI tools—ChatGPT, Google AI overviews, Google Gemini, and Perplexity—by submitting 100 news-related queries across these platforms and analyzing over 2,500 links generated by their responses.

The findings revealed that ChatGPT and Gemini did not cite content from the BBC, as the broadcaster has prohibited the bots from accessing its news articles. However, Google AI overviews and Perplexity included BBC journalism despite the broadcaster’s disapproval of their tools utilizing its content.

The analysis also discovered that sources like the Telegraph, GB News, the Sun, and the Daily Mail accounted for less than 4% of citations in responses generated by ChatGPT. In contrast, the Guardian, which has a licensing agreement with ChatGPT’s parent company, OpenAI, was referenced in nearly 60% of cases. Moreover, the Financial Times, which also collaborates with OpenAI, was noted frequently, indicating a potential disparity in the representation of sources.

The implications of Google’s AI summaries at the top of search results have adversely impacted click-through traffic for publishers, leading to reduced revenues, as many users engage with the summaries without further exploring the original articles.

The report raises important questions regarding the influence of financial arrangements between AI companies and news outlets on the answers generated. It warns that should licensed publications be prioritized in AI responses, smaller and local media organizations might be disproportionately marginalized, given their limited chances of securing AI partnerships.

Although the IPPR acknowledges that licensing contracts could partially replace advertising revenue, it cautions that they may not sustain a balanced news ecosystem. This reliance could inadvertently tie news organizations to tech giants, exposing them to financial risks should copyright laws weaken.

To offset these challenges, the IPPR advocates for public funding aimed at fostering new business strategies for investigative journalism and local news, which face sustainability threats due to the ascent of AI-generated content. It also urges support for the BBC to “innovate with AI.”

In response to these assertions, an OpenAI spokesperson stated that the organization is dedicated to supporting quality journalism. They explained, “When ChatGPT searches the internet, it draws from a comprehensive range of publicly available sources relevant to the inquiry and displays citations and direct links. This not only aids users in verifying information but also drives traffic to publishers.”