This week, it’s been reported, news media lobbyists are heading back to the Hill to make their case for what they describe as “legislation to protect and support quality journalism.” In recent years, this effort has largely centered on the Journalism Competition & Preservation Act (JCPA), a bill that would create an exemption to antitrust law, which normally prohibits industry collusion on pricing and other business terms. It would allow certain news organizations to band together into “joint negotiating entities” to negotiate for payment from dominant digital platforms for “accessing” (i.e., crawling, indexing, and/or displaying snippets or thumbnails of) their content. In each of the past two years, lobbyists have tried to get the JCPA included in a year-end, must-pass defense spending bill. While we don’t have any evidence that the JCPA will find its way into such legislation this year, we thought this might be an apt moment to post a refresher list of the reasons we oppose this legislation. (We’re not alone: a wide range of groups that are normally at odds over tech policy and many other issues have joined forces over and over again to oppose this bill.) Our impressions here are based on the last iteration of bill text we saw in 2024.
The last time news lobbyists headed to the Hill to push for the JCPA, it was partly in response to the failure of two California state bills designed to extract monopoly rents gained by Google and Meta for allocation to publishers. The bills were both based on the reasoning that the dominant platforms have used their monopoly positions to steal traffic and advertising revenue from publishers. In the end, California Gov. Gavin Newsom shelved both bills, and instead announced a “News Transformation Fund” to be voluntarily co-funded by Google and the state. The fund’s design favored print and digital outlets, small outlets, under-served markets, and the retention (if not hiring) of journalists. Lobbyists for the JCPA claimed the failure of the California bills proved the need for federal legislation, and that if publishers were empowered to negotiate collectively (without breaking antitrust laws), they would have more leverage with the platforms.
Another factor in play at that time was the liability decision in the Google search case. The judge decided that Google violated antitrust law in its conduct in the search market and leveraged its power in the search market to dominate the search text advertising market. This was positioned by news industry lobbyists as proof that Google is a monopolist with an illegal business model and therefore, owes publishers recompense. While we agree that Google is a monopolist, it does not follow that publishers are free to extract rents for links.
What’s New?
There have been some new developments since then that may trigger new advocacy for the JCPA. First, publishers were just handed what we agree is a crushing disappointment of a judicial decision regarding the remedies in the Google search case. The judge missed many opportunities to implement meaningful structural and behavioral remedies to Google’s monopolistic practices. Most relevant to the JCPA, he declined to adopt either of the Department of Justice’s recommended remedies that would have benefited publishers: prohibiting agreements that give Google exclusive access to publisher content, or allowing publishers to opt out of being crawled by AI. Despite our agreement that the judge’s decision in the search case was a whiff, we do not feel the JCPA is a good solution to the challenges to publishers brought about by AI.
The second development: the first two judicial decisions in lawsuits by publishers against AI firms both affirmed that under current law, AI training is highly transformative and a fair use of copyrighted content (at least in the situations analyzed in the cases). In other words, AI firms do not infringe on publishers’ rights when they use copyrighted news content to train their models. This is not the outcome publishers wanted, and it may lend tailwind to the idea, first framed in early 2024, that the JCPA will create a stream of compensation from AI companies back into news because the bill’s language about “accessing, crawling, and indexing” news content covers its use to train large language models. We disagree with this argument, but there may be an appetite for it in Congress. There have already been sympathetic hearings and more to the point, three Senate bills introduced calling for transparency, at a minimum, and restrictions, at a maximum, regarding the use of copyrighted content to train large language models. Early in September, Sen. Josh Hawley (R-MO) also argued there should be property rights assigned to certain types of data, and legal liability for companies that use it to train their models. News publishers may also point to the settlement in Anthropic’s class action copyright lawsuit with authors whose books were allegedly pirated for use in Anthropic’s training data. It implies that “access” to content for AI training warrants compensation.
Here, updated for the AI age, is a list of the reasons we oppose the JCPA.
The JCPA upends copyright law, and not for the better.
As we’ve noted before, the JCPA as drafted could be interpreted by courts to implicitly expand the exclusive rights that news publications enjoy in their material, beyond what any copyright owner has ever enjoyed. (For publishers, this is the point.) This would represent a major shift in copyright law (not to mention the nature of the internet, which is fundamentally built on the user freely linking to content). If interpreted by courts in that way, the JCPA could prevent internet users from linking to or sharing news articles on digital platforms and websites without some sort of payment. The bill itself also allows publishers to deny the platforms access to their content, which may restrict users from posting links to news stories. This would further limit the public’s access to information online. If applied to the use of copyrighted content for training of AI models, the JCPA upends fair use, the legal doctrine that allows limited use of copyrighted works without permission. Ironically, journalists themselves depend on fair use for criticism and commentary, news gathering and reporting, republishing source material, illustration, historical reference, and documenting claims.
In fact, the JCPA is based on a premise the U.S. Copyright Office rejected.
Back in 2021, when publishers’ frustrations with technology firms were still focused mostly on search and social media platforms, a few members of Congress sent a letter to the U.S. Copyright Office asking for “a study to evaluate the effectiveness of current copyright protections for publishers in the United States”. It was a sign of consistent Congressional support for the needs of news publishers, in particular, to confront technological disruption. The letter pointed as precedent to a European Union directive establishing “ancillary copyright” protections for press publishers. In June of 2022, still six months before the explosive introduction of ChatGPT, the Copyright Office issued its report, “Copyright Protections for Press Publishers.” It concluded that the publishing industry already benefits from significant protections under existing law and does not require any new copyright protections. Public Knowledge provided public comments for the preparation of that report and supported its conclusions. We noted then, as we do now, that restricting linking and sharing of information through changes to copyright law, like the European Union has done, undermines the open nature of information on digital platforms and invariably favors and entrenches the largest players. A subsequent Copyright Office report specifically addressing AI and copyright affirmed that existing copyright law – including the fair use doctrine – can be applied to address the new questions brought about by generative artificial intelligence.
The JCPA discourages content moderation, even if it’s consistent with platforms’ terms of service.
Platforms’ content moderation policies and processes have evolved a great deal since the last Congressional consideration of the JCPA. And so has, apparently, the Republican party’s view on the desirability of platforms moderating political or “hate” speech. But one thing remains true: an amendment to the JCPA made in late 2022 discourages or outright prevents platforms from using content moderation to support their community standards or terms of service. The amendment nullifies the antitrust immunity offered to journalism providers if the case could be made that any of their discussions with platforms were about content moderation, or that platforms discriminated among publishers based on “viewpoint.” Platforms will play it even safer to avoid being accused of “viewpoint discrimination” by erring on the side of moderating less content. That means users will see more harmful false information, extreme content, and hate speech online.
Without changes, the JCPA doesn’t do enough for journalists.
Last thing we knew, there were discussions about changing the language of the JCPA to include a requirement that 70% of proceeds from negotiations or arbitration go to newsrooms, and that there is transparency on how all proceeds are spent. If those changes are made, the bill may earn the support of the NewsGuild, the dominant union of journalists and media workers in the United States. (Both of these provisions would be derived from the [failed] Californian version of the bill, which earned the support of NewsGuild units in California.) The NewsGuild’s support may be less influential in the bill’s passage than it was before Republicans had the majority in the Senate and the House. That said, we still feel that any kind of negotiation between platforms and publishers will primarily benefit the biggest media conglomerates, many of which are owned by financially-motivated hedge funds. Which brings us to…
The JCPA primarily benefits the largest publishers.
Lobbyists for the JCPA have pointed to a requirement in the bill that publishers with more than 1500 employees are not eligible to join a negotiating entity as proof that it benefits small publishers. But as we pointed out in a previous post, the publisher employee cap excludes only three elite east coast newspaper publishers in the United States from joining negotiating entities, and it doesn’t apply to broadcasters at all. So Sinclair Broadcasting Group, with approximately 13,000 employees and $2.59 billion in profit in the first quarter of 2022 alone, can participate in a negotiating entity. The JCPA as we saw it last also doesn’t allow news organizations that have been in business for less than a year to enter a negotiating entity, and it excludes news organizations that earn less than $100,000 per year. In other words, it is designed to disadvantage the smallest publishers. That’s why organizations representing local independent online news organizations, like LION Publishers, have traditionally opposed the bill. So have organizations representing the Black Press.
The JCPA increases publishers’ reliance on dominant platforms.
At Public Knowledge we have consistently written about the need to center news policy on the information needs of communities – not on protecting legacy news models. When corporate interests, whether those of tech or media companies, begin to erect walls, levy tolls, or make backroom bargains, it will be the public that loses. In that sense, the JCPA overlooks important lessons we’ve learned from similar legislative actions in Australia and Canada.
In late 2024, a Joint Select Committee on Social Media and Australian Society issued a report noting that their News Media Bargaining Code, the model for the JCPA, was based on fundamentally incorrect assumptions about the value of news to dominant platforms. Meta had signaled their intention not to renew commercial agreements with publishers in Australia, and threatened to break news links in Australia if they are “designated” and therefore forced to forge agreements with publishers under the Code. Australian regulators backed away from the Code. Similarly, upon passage of Canada’s Online News Act, Meta blocked news links in Canada. About a year in, online traffic, engagement, and revenue plummeted for publishers – and no one seems to have noticed the absence of news online in Canada.
The JCPA is inconsistent with the new administration’s plan for AI.
While we have considerable concerns with the Trump administration’s AI action plan, we do agree with its implicit support for fair use and for allowing courts to decide on the use of copyrighted content in model training based on the distinctive facts in each case. We also appreciated the President’s own remarks about the need to protect the rights to read and learn. The JCPA is incompatible with these aspects of the administration’s plan for AI development.
It would also be hypocritical if Republicans supported the JCPA given the Trump administration’s strong comments, especially via the FTC, about how dominant platforms “censor through market power” and “elites collude,” including through their trade associations, to decide who will be heard. We find this line of thought off-base. Regardless, the JCPA would increase platforms’ influence over online publishers and grant news organizations the cover to collude to set the pricing, terms, and conditions under which platforms have access to speech.
There are better options available to legislators to protect publishers from the impacts of AI.
We are genuinely sympathetic to publishers’ concerns about the impact of AI on their business models, and have other solutions to offer. First, as we have said before, we cannot rely solely on judges to forge antitrust solutions to mitigate the monopolistic practices of Big Tech and their impact on publishers. Congress must pass legislation that breaks up the consolidated, abusive power of online platforms to pick winners and losers in the race for online visibility. Congress must also pass legislation that mandates transparency and prohibits self-preferencing. Examples of existing proposals to address this include the Advertising Middlemen Endangering Rigorous Internet Competition Accountability Act (AMERICA Act) and the American Innovation and Choice Online Act (AICOA). Another option is to put into law one of the remedies the DOJ framed in the Google search case: requiring that dominant firms like Google, Apple and Microsoft sever their AI crawlers from their search crawlers. This would support statutorily enforcing a set of self-identification standards via a unique identifier for a bot crawling content.
In addition, we have recently framed a set of middle-ground solutions for discussion that are designed to preserve the benefits of the fair use doctrine for creativity and free expression while still preserving incentives for publishers to produce timely content. These include extended collective licensing, self-identification standards for bots and crawlers, and text and data mining exceptions for public interest uses. For more about these solutions, see “Is There a Middle Ground in the Tug of War Between News Publishers and AI Firms? Part 1: Framing the Problem” and “Is There a Middle Ground in the Tug of War Between News Publishers and AI Firms? Part 2: Framing Solutions.”
For all of Public Knowledge’s resources about the JCPA, see our resources page.