The Clash Between AI and Journalism - Japan Also Involved: The Lawsuit by Asahi and Nikkei Highlights the Risks of "Generative Search"

1) It's Not Just "Copyright": The Focus Has Shifted to "Trust"

Search engines vs. newspapers, SNS platforms vs. publishers—the legal battles in the media industry have traditionally been against giant platforms controlling the "distribution channels." However, with the advent of generative AI, the axis of conflict has begun to twist. Instead of "linking and sending" news, AI now "answers on the spot." Often, these answers are confidently incorrect, and they cite sources like "according to such-and-such paper." Thus, the focus of disputes is shifting from mere reproduction and reprinting to **the handling of sources and brand damage (trust cost)**. NDTV Profit

NDTV Profit illustrates this very change. Newsrooms and authors are beginning to legally pursue not just "our work was used for training," but also the danger of "misinformation being spread under our name." NDTV Profit

2) Spark ①: Prominent Investigative Journalists Join the Fray—Pressuring AI Companies with "Individual Lawsuits"

A symbolic case is the lawsuit filed by John Carreyrou, known for exposing the Theranos scandal. Along with several authors, he is suing xAI, Anthropic, Google, OpenAI, Meta, Perplexity, and others, claiming they "used copyrighted works (books) for training without permission." Furthermore, this lawsuit is designed as individual lawsuits rather than a class action, highlighting the concern that "in class actions, companies can 'settle cheaply.'" Reuters

In the background, Anthropic agreed to a settlement of approximately $1.5 billion in a class action over the "use of pirated books" (debates over distribution and attorney fees continue). Some authors perceive this as insufficient deterrence and aim for stronger conditions by "fighting individually." Reuters

3) Spark ②: NYT's Two-Front Strategy—"Copyright" and "Trademark/Misrepresentation"

Another focus emphasized by NDTV Profit is the movement involving The New York Times (NYT). NYT raises concerns that generative AI confuses readers by using its articles as material while providing incorrect summaries or answers, thus damaging brand value (trust). The key here is expanding the issue beyond mere copyright infringement to consumer misrepresentation and trademark damage (in the U.S., under the Lanham Act)—viewing "hallucinations" not as a technical flaw but as commercial damage. NDTV Profit

It is also reported that NYT has sued Perplexity for unauthorized mass copying and displaying of articles. As the "answer screen" of generative AI becomes the main battlefield for news consumption, media will use "misinformation and attribution" rather than "presence of links" as their weapon. The Guardian

4) Spark ③: "Summaries Kill Clicks"—Penske Sues Google's AI Overviews

The battlefront is also expanding to search. Reuters reported that Penske Media, which owns Rolling Stone, Billboard, and Variety, has sued Google over its AI Overviews (AI summaries appearing at the top of search results). Penske's argument is simple: if summaries are complete on the search screen, readers won't visit the article itself. As a result, revenue from ads, subscriptions, and affiliates is damaged, shaking the foundation of news production. Reuters

What's important here is not just that generative AI "steals/learns content," but that the very UI design of search structurally reduces traffic to media. This area easily connects copyright disputes with competition policy and market dominance (monopoly) discussions. Reuters

5) Japan Also Becomes a Stakeholder: Asahi & Nikkei vs. Perplexity, Focusing on "Technical Evasion" and "Incorrect Answers"

NDTV Profit mentions that in Japan, Asahi Shimbun and Nikkei have sued Perplexity in the Tokyo District Court. Financial Times also reported that both companies are concerned about unauthorized copying and storage of articles, evasion of technical restrictions, and the damage to credibility when misinformation is presented under their name. Financial Times

What is characteristic in Japan is that not only copyright but also **unfair competition (free-riding)** logic easily comes into play. AI's answers appear as "quotes," but in reality, they siphon off the reader's news experience to the platform side. Moreover, if errors are mixed in, the media's credibility is used as collateral without permission. This becomes an issue of brand security more than "whether it was reprinted." Financial Times

6) BBC Also Issues Warning: Demands "Stop, Delete, Compensate" from Perplexity

In the UK, Reuters reported that the BBC is demanding Perplexity to stop scraping content, delete stored data, and propose compensation, while considering legal action. Perplexity has pushed back, arguing that the other side lacks understanding. Reuters

The reason why the public broadcaster BBC is stepping in so deeply is likely due to a sense of crisis that "the credibility of public information" is being damaged by AI summaries, not just revenue. When AI handling news wavers in both "accuracy" and "source," the cost ultimately falls on societal decision-making. The Guardian

7) Why "Hallucinations" Are No Longer a Laughing Matter—The Source Issue Highlighted by EBU/BBC Research

Backing this concern with data is research involving the EBU (European Broadcasting Union) and the BBC. Reuters reported that in AI responses to news-related questions, serious sourcing issues such as omissions, misguidance, and incorrect attribution were observed in about one-third of cases. Reuters

While the incorrect answers themselves are painful, what media fears more is the "spread of errors under their name." The more generative AI speaks in an "authoritative tone," the less readers scrutinize the source. As a result, a media outlet's trust can be significantly damaged by a single misattribution. The shift from "copyright to brand damage" pointed out by NDTV Profit has its basis here. NDTV Profit

8) Yet "Transactions Continue": The Reality of Simultaneous Lawsuits and Licensing

Interestingly, while lawsuits are increasing, so are licensing agreements. For instance, Reuters reported on a "landmark agreement" between Axel Springer and OpenAI. Reuters

Furthermore, Financial Times reported that NYT has entered into an AI-related agreement with Amazon. Financial Times
AP also touched on the licensing agreement with OpenAI amid ongoing lawsuits involving NYT and others. AP News

Thus, the market is split into two.

• The camp aiming to establish rules in court (deterrence and precedent-setting)

• The camp securing negotiation power by monetizing through contracts first (short-term defense and revenue)

However, this "transaction" is not a permanent solution. The core issues of how much learning is allowed, what citation UI or source display is mandated, and who bears responsibility in the event of misinformation remain unclear. NDTV Profit

9) Moving Towards State Intervention: India's "Comprehensive License" Proposal

Taking it a step further is the design of a national-level system. In India, it is reported that the DPIIT (a department of the Ministry of Commerce and Industry) has proposed a framework akin to a **"comprehensive compulsory license + royalty collection agency"** for AI learning. The idea is to allow companies broad access to copyrighted works without individual permissions, in exchange for centralized collection and distribution to compensate rights holders. The Economic Times

This approach is faster than settling each case through litigation. However, it also raises new issues such as "handling of opt-out rights," "fairness of distribution," "transparency," and "application to foreign companies." The world is simultaneously experimenting with both the "court-drawn line route" and the "systematic processing route" as of the end of 2025. The Economic Times

10) Reactions on Social Media: Broad Sympathy but Divided Opinions

From here, we will organize the "trends" of reactions based on discussions on public social media (such as Reddit), noting that these represent only a portion of community posts and not the entire public opinion.

(1) "Finally, the Main Issue is on Fire" Camp
In Reddit threads discussing Carreyrou's lawsuit, reactions like "this battle was destined to go to court" and "the beginning of a larger legal struggle" are observed. While acknowledging the convenience of AI, there is frustration over the commercial use of training data while its handling remains "ambiguous." Reddit

(2) "Lawsuits Won't Stop It, It's Ultimately About Contracts and Technology" Camp
On the other hand, there is a persistent view that even if lawsuits increase, the use of AI itself won't stop. Therefore, the discussion tends to focus on the combination of technology and contracts as a realistic solution, such as strengthening UI design for source display,