Perplexity's moment
Buying Perplexity could finally give Apple AI credibility. Plus: 3 copyright rulings on fair use, Axel Springer puts a ring on AI, and more.
Perplexity may not create foundational AI models, but it's been punching above its weight for most of its existence, quickly becoming the go-to AI search engine for many. With success comes attention, and now it may be at a crossroads, with many speculating that an exit via a Big Tech acquisition might be its best option. The media is playing a role here, turning up the legal heat over alleged copyright violations. What should its future be?
More on that in today's column, but a quick reminder that my monthly live AI class, AI Quick Start, is back. Tailored for journalists and PR professionals, the class will give you both foundational knowledge on how to approach AI for your job and several practical use cases you can start applying right away. Next class is July 17—more details here 👇
A MESSAGE FROM THE MEDIA COPILOT
Let’s be honest. The hype around AI is ridiculous, the tools can feel overwhelming, and most advice just doesn't resonate with professionals doing real journalism or PR work.
That’s why I’m hosting a focused, one-hour AI class specifically for journalists and PR professionals. It cuts through the noise to highlight what really matters: how you should think about and use AI.
This isn't a rapid-fire rundown of tools. It’s a fundamentals-first session designed to help journalists and comms experts adapt and thrive in an AI-driven media landscape.
You’ll walk away with:
A clear, practical overview of the current AI landscape
Guidance on the mindset needed to effectively leverage AI
Valuable, actionable use cases tailored for journalism and PR
The details:
🗓️ July 17 at 1 p.m. ET
💸 $49 — an incredible value for essential insights
🎁 Included: class materials, actionable prompts, and a special gift
Why this might be the best time for Apple to buy Perplexity
The buzz around Perplexity has been deafening lately. It's apparently an acquisition target of two of the biggest names in tech, but it's also attracting more attention over copyright concerns. If it's looking for an exit, this might be the best time.
All the recent chatter started when Bloomberg reported that Meta tried to acquire Perplexity before it instead invested $14 billion in Scale AI in order to poach its CEO, Alexandr Wang. Just a few hours later, Bloomberg reported that Apple had internal talks about buying Perplexity, though they're only in the early stages. The new report puts some meat on comments that Apple executive Eddy Cue made in May during one of Google's antitrust trials. Various commentators praised the idea.
Right around the same time, the BBC threatened to sue the AI search engine, claiming that it reproduces content "verbatim," violating copyright law, in a letter to Perplexity CEO Aravind Srinivas. The news org says Perplexity's AI summaries damage its reputation since they're sometimes inaccurate and fall short of BBC editorial standards.
The legal pressure is on
If there's truly no such thing as bad publicity, Perplexity has had a very good week! But neither story bodes well for the company in the long term. One story suggests its days as an independent entity are numbered, and while the BBC's action is a threat, not a lawsuit, it does feel as though whatever grace it's been given by the media community is starting to wane.
The BBC's move isn't the first time Perplexity has been the subject of legal attention from a publisher. News Corp skipped the threat phase and full-on sued Perplexity last year. The difference is News Corp also has a lucrative deal with OpenAI to license its content, and accused Perplexity of trying to effectively backdoor a license via its Publishers' Program, which shares ad revenue with partners (the revenue is nascent). The BBC has so far not signed any licensing agreements with tech companies.
The BBC's stance seems to be more principled than economic, but it amounts to the same accusation: taking content and then building a product that competes with that content—with no compensation.
This is where I wonder if Perplexity can survive this growing legal onslaught. Broadly, the whole idea of using AI to summarize real-time web content (often called RAG) is playing out much more like syndication than search. Several AI-based knowledge services, from Dow Jones Factiva to ProRata are proactively seeking licensing for content. That's rapidly becoming a standard practice, and companies that rely on the old paradigm of the search era—scrape everything and anything, then do whatever you want with it—are starting to look behind the times, especially now that there's a lot of good data that shows AI summaries destroy referral traffic.
It seems to me that for a company like Perplexity, which doesn't build foundation models and bases its entire business model on AI search, has a lot more to lose if these coming legal battles decide licensing is required to use content for search/RAG applications (although that's far from a guarantee—see the Chatbox below).
Apple's need for an AI savior
Given the legal quicksand Perplexity is standing in, an early exit might be the best option, and Apple desperately needs an AI boost. Siri still doesn't have the AI features promised over a year ago, and after an internal reshuffle, Apple has pushed back plans for advanced AI functions (like Siri adapting to your personal data and preferences, and being able to perform in-app actions) until 2026. Buying Perplexity would instantly put Siri on solid footing by enhancing it with a proven AI search engine.
True, Perplexity doesn't actually make frontier models, which might make it a less attractive acquisition target. But owning its own model wouldn't necessarily fix Apple's reputational problem in AI. The fact is no one in Silicon Valley thinks of Apple as a great place to do cutting-edge AI research, and that would still apply even if it acquired Anthropic or whoever. And you can't become a premier AI lab without the right talent.
Perplexity is also rising in popularity. Usage continues to rise: media sites saw bot activity from Perplexity increase 359% just this past quarter, compared to 56% generally, per data from TollBit. If Apple doesn't acquire the company now, it could get considerably more expensive in the coming months.
It all looks good on paper, until you remember Perplexity's legal problems. The complaints from News Corp and the BBC still apply, even if Apple owns the company. And if the courts end up judging AI search engines harshly, Apple will look pretty stupid for paying tens of billions for a company with a fundamentally flawed business model.
What's working in Perplexity's favor is time. Apple's need to turn its AI ship around is urgent, and ironically, Perplexity is having a moment thanks to all the rumors of tech companies looking to cut a check for the company. If Perplexity takes that check, it would be a whole new AI race, and maybe Siri could finally put "Here's something I found on the web..." to bed permanently.
The Chatbox
All the AI news that matters to media*
The copyright pendulum swings back
In three different courts, AI companies scored minor victories in the ongoing legal war over copyright. As reported by Wired, the most notable was a federal judge declaring that Meta training its AI models on books, including one written by comedian Sarah Silverman, did not violate copyright. The judge left open the possibility of future lawsuits being successful, though, saying that the plaintiffs simply didn't establish that Meta had caused them financial harm. In a separate case, AI Fray reports another judge said Anthropic's use of books in training data fell within fair use, and Getty Images dropped some of its key copyright claims in its UK case against Stability AI, makers of Stable Diffusion, per TechCrunch. None of the rulings on its own were decisive on the subject of whether AI training constitutes a copyright violation, but the winds appear to be shifting after earlier cases favored content creators over Big AI.
Axel Springer makes its AI pivot official
Axel Springer's declaration that "the business model of maximizing clicks and advertising is over" might solicit a "duh" from anyone who's been paying attention, but the fact that it's saying so out loud in its official strategic update to investors represents a fundamental shift in the business of media. The German publisher behind Politico and Business Insider plans to double its value over five years by pivoting from traffic-driven content to AI-powered journalism and direct user relationships, explicitly reducing reliance on search engines and social media platforms. The move makes sense given the rapid rise of AI search engines and Google's push toward AI Mode. However, exactly how that will translate into more dollars absent clear business models still hasn't been worked out. Everyone will be looking to companies like Axel Springer to not just pivot, but architect where we're pivoting to.
Oregon's bill forcing Big Tech to pay for news fails to pass
The Oregon bill that would have charged tech companies $122 million annually for sharing news content is dead, reports Oregon Public Broadcasting. While Senate Bill 686 had backing from major Oregon news organizations, Meta's preemptive threat to pull news from its platforms—as it had done in Canada—likely influenced the narrow 15-14 defeat. The bill's failure exposes how platform companies can effectively veto legislation by threatening to cut off distribution that newsrooms have become dependent on, even as those same outlets struggle with mass layoffs and declining revenue. California's recent example, where Google slashed its promised newsroom funding by two-thirds after striking a deal, suggests that even "successful" negotiations with tech giants may offer false hope. (AI-assisted)
From prose to TikTok in three clicks
Sophiana wants to bridge the cultural divide between word-driven journalists and the video-first audiences they're losing on social platforms, reports Nieman Lab. The app, developed by UK journalist Sophia Smith Galer, uses GPT-4o to transform written stories into videos tailored to platforms like TikTok and Instagram. Here's the uncomfortable truth behind the tool: many journalists look down on platforms like TikTok because they prioritize engagement over depth. Smith Galer argues that journalists should "better saturate" platforms to provide a counterweight to misinformation rather than avoid them entirely. In any case, Sophiana is a step toward a world full of "liquid content," where AI remixes and reformats content to match the user's preference. At some point journalists may not have a choice but to simply let the bots do what they want with the facts they uncover. (AI-assisted)
*Some items are AI-assisted. For more on what this means, see this note.