The term ‘artificial intelligence’ – at least in the information context – is an oxymoron. “Artificial” may be accurate, but when it comes to research on current events, AI adds nothing intelligent except a natural-language search engine. And even then, the users can only rely on the intelligence of the source material AI scrapes to generate its output.
Two developments today highlight that clear conclusion. Let’s start with the New York Times, which has filed a lawsuit against an AI platform that allegedly uses the Paper of Record to feed its AI output. Not surprisingly, the NYT’s lawyers take exception to commercial reuse of their material without compensation or fair-use purpose, and want the new Perplexity platform to either pay up or shut down.
They’re not the only media outlet ‘perplexed’ by Perplexity’s arrogance, either:
The New York Times claimed in a lawsuit on Friday that its copyrights were repeatedly violated by Perplexity, an artificial intelligence start-up that has built a cutting-edge internet search engine.
The Times said in its lawsuit that it had contacted Perplexity several times over the past 18 months, demanding that the start-up stop using the publication’s content until the two companies negotiated an agreement. But Perplexity continued to use The Times’s material.
The suit, filed in federal court in New York, is the latest in a growing legal battle between copyright holders and A.I. companies that includes more than 40 cases around the country. In August 2024, Dow Jones, owner of The Wall Street Journal, The New York Post and other publications, made similar claims in another lawsuit against Perplexity.
Say what you will about media outlets, but regardless of the quality of their content, they own it. It’s protected by copyright, and it’s their property. We excerpt material from these same media outlets – in some cases, much to their chagrin – but we do so within the boundaries of “fair use.” We excerpt it with links and attribution, and we only excerpt as much as we need to provide commentary, context, and criticism. To use this material as our own output, we would have to license the material from the NYT and all of the rest of these platforms. Otherwise, it’s just theft.
This is not a new concept, even in the digital media world. The search engine for Bing feeds Microsoft’s MSN platform for many media reports, including from some of the outlets listed above as well as Bloomberg, Reuters, and others, whose content MSN publishes in full. Microsoft has licenses from these outlets for that purpose, and likely pays a lot of cash for those clicks.
AI engines in this space basically act as search engines too, only without the licenses that allow for republication – or so the NYT and these other outlets allege. They are using natural-language search engines to produce output like Bing’s normal search does for news, only apparently stripping out the attribution (and licensing) to claim the output as original. That’s precisely what the NYT alleges in its new lawsuit:
The suit accuses Perplexity of violating The Times’s copyrights in several ways, most notably when the start-up’s search engine retrieves information from a website or database and uses that information to generate a piece of text and to respond to queries from internet users. That would not be a fair use of that material, the suit claimed, because Perplexity grabbed large chunks of the publication’s content — in some cases, entire articles — and provided information that directly competed with what The Times offered its readers.
“Perplexity provides commercial products to its own users that substitute for The Times, without permission or remuneration,” the suit said.
The Times also accused Perplexity of damaging its brand. In some cases, the suit said, Perplexity’s search engine made up information — what A.I. researchers call “hallucination” — and falsely attributed that information to The Times.
If that’s the case, Perplexity will have more problems than just licensing. It also points up a problem seen in other AI contexts, notably in the legal industry, where artificial “intelligence” turns out to be fatally stupid. More than one trial attorney has been burned by AI-generated motions citing cases that do not exist, meaning attorneys submitted false information to the courts. If the NYT can prove damage to its reputation, then all sorts of possibilities open up for extensive damages.
Of course, the NYT would have to have a positive reputation to damage. But I digress.
Meta has decided to approach its AI information engine in a more principled manner. They also announced today that they will partner with a surprisingly wide range of media outlets for licensed access to feed into its AI engines:
Meta on Friday said it has struck several commercial AI data agreements with news publishers ranging from USA Today, People Inc., CNN, Fox News, The Daily Caller, Washington Examiner and Le Monde.
Why it matters: Meta backed away from compensating news companies several years ago, but has resumed paying for editorial content as it invests more in its AI chat tools that require real-time access to verified news.
Zoom in: The new agreements allow Meta to access content from its partners to provide real-time answers to user queries about news and current events in its Meta AI chatbot.
Will that solve the problems of AI “hallucinations”? Probably not, nor will it solve the Garbage In Garbage Out issue that some of these outlets will create. At least Meta has decided to include more conservative-leaning outlets like Fox, Washington Examiner, and the Daily Caller for its source material, however. That’s a notable shift for Meta and Mark Zuckerberg.
