Journalism faces unprecedented challenges in 2025 as the intersection of artificial intelligence, content creation, and business models reaches a critical juncture. As Google and Meta continue to pivot away from news while introducing new generative AI search and chatbots, the limited revenue and access to audiences that publishers were able to obtain from referral traffic is set to decline further, meaning that adaptation is not optional — it’s existential.

The great content divide

The valuation and protection of journalistic content in an era of aggressive AI-driven content extraction is taking place against a backdrop of increasing integration of AI in the newsroom and operations. After a flurry of partnerships and licensing deals with the most prestigious and content-rich news brands over the past year, Big Tech firms Google, Meta, Microsoft, and (Microsoft-backed OpenAI) are unlikely to pursue as many in the year to come as they continue to claim fair use for text and data-mining amid a slew of lawsuits that will take years to resolve. This strategic co-optation of publishers with the biggest and highest quality portfolios — which also happen to be those with the resources to sue — will create a stark divide in the media ecosystem, where only the largest publishers with guaranteed content pipelines will secure AI partnerships with the Big Tech firms and AI unicorns. This will leave the long-tail of smaller, regional, local, ethnic, investigative, and specialized outlets struggling to adapt to an environment in which referral traffic from search and social media continues its precipitous decline amid increasing competition from AI-generated content farms and AI-curated news apps.

However, there is also an opportunity for those newsrooms left out of deals — the “long tail” of journalism. AI companies will need to solve the problem of declining access to quality data — a trend that is likely to become more acute as publishers restrict the unfettered crawling of their websites by AI bots and experiment with new ways to license and monetize their archival and real-time content. There is also a growing array of AI startups that are consumer-facing or aim to provide business-oriented services that rely on access to relevant, timely, and reliable content and can’t afford to flout copyright or offer indemnity to their customers, like the Big Tech firms do.

From RAGs to riches? The content valuation challenge

Publishers are entering this complex new terrain of content valuation with limited information on how to establish the value of their journalism in this new AI-driven ecosystem. This means that 2025 will be the year of learning and experimentation. Publishers will figure out how to distinguish between training and inference data, and why retrieval augmentation generation (RAG), a process that retrieves relevant information from external sources to supplement large language models, could offer more diverse and lucrative revenue streams than selling off their archives to train large language models. They will need to experiment with a range of licensing solutions that aim to connect content industries with AI models. The news industry will be forced to develop nuanced pricing models for different content use cases, with particular attention to archival content, original journalism, and local reporting.

This represents a crucial pivot from previous passive approaches to content licensing, making it imperative to learn from other industries like music, which has a more nuanced and dynamic approach to licensing that includes rights and compensation for musicians and composers, not just corporate content owners.

Adaptive partnerships for the AI data marketplace

2025 will see significant experimentation between publishers and AI companies, from those run by former Big Tech news execs like Tollbit and HumanNative to others like ProRata, Sphere, and Emergent Methods that have fewer ties to the platforms that eviscerated the business models of digital journalism. This competitive innovation landscape creates unique opportunities for publishers and independent journalists to explore diverse AI licensing models and see what works for them. With so many competing startups vying to serve this emerging market, there will be lots of opportunities for publishers — and in some cases journalists who have their own newsletters or websites — to experiment and try out the different services. Which means that investing in technical expertise, forming collaboratives to learn and perhaps even share technical resources, and collective approaches will become increasingly important.

2025 will be defined by journalism’s strategic recalibration in response to AI’s disruptive potential. Success will depend on publishers’ ability to understand their content’s value, develop technical sophistication, and navigate a complex, rapidly evolving technological landscape all while continuing to produce journalism in an increasingly challenging and hostile political environment.