2025: The year we stop talking about “AI”

This article was original published as part of Nieman Lab's Predictions for Journalism, 2025.

We’ve gotten through the first part of the hype cycle on artificial intelligence. Just over two years since ChatGPT launched and raised mass awareness of this new technology, and billions of dollars of venture capital later, most people are aware of AI and a good number have tinkered with it.

So where do things go now? People don’t care about the infrastructure of your website or mobile app. They don’t care if you’re using Python or Swift, Postgres or MongoDB, Nvidia or Google processors. They don’t care if you’re using generative AI. They care about what it does for them and if it’s actually useful.

That’s where things are heading in 2025. Companies have started experiments, built internal tools, and even created their own ChatGPT-like chatbots. It’s now time to truly tap the potential of this technology or else be run over by those who do. It’s time for publishers to build the native AI experiences that power new types of media products.

Over the past year, we’ve been working with journalists and technologists around the world at hackathons and workshops to envision this future. We’ve asked people to imagine they are in the year 2030 launching a new media startup, or that they are gifted $5 million from generous donors to fund an AI project. We’ve given them fictional promotions to head innovation labs at media companies, charged with creating the new AI-based product that will become their main revenue source. Hundreds of participants have given demos, created pitch decks and organization charts to visualize this future.

It’s hard to think about new media products from first principles, discarding all preconceptions. We’ve tried to push people to get rid of all legacy baggage about what a “news story” should be, to not hold anything sacred or untouchable.

In getting people to make this mental switch, there are two big trends driving change now in the way media works. First, AI makes personalization possible to the individual level, shrinking any notion of mass media. Every person can see what they want in the way they want it. Second, legacy media brands are experiencing a crisis of trust, and personality-driven media is taking its place.

If we put these two trends together, we start to see glimmers of new media formats and product features. Media takes on the characteristics of algorithmic feeds and conversational interfaces. Everything that we know about a consumer is brought into play, even their biometric data. AI can help with the challenges that have made engagement features too expensive or risky in the past, pitching in to moderate community-generated content, or matching people with shared interests to have a constructive conversation.

To get started, media companies who succeed will foster this type of thinking with their own staff. They’ll start by engaging with their community, and make them a part of the product development process. This has the benefit of also recruiting the future community evangelists for whatever eventually launches. Newsrooms will experiment and run internal hackathons to create prototypes. The companies who are succeeding will launch test products with “beta” labels and be open about how things work and might not work, and always be asking for more outside feedback. Not moving at all, or even moving too slowly, is a recipe for obsolescence.

As we head into the new year, success will come from remembering what makes media powerful in the first place: its ability to inform, connect and inspire. The best AI-powered media products will create spaces for genuine human connection, foster understanding across divides, and help rebuild trust. Rather than worrying about AI taking away jobs or isolating us further in society, the best media products will help us connect and be more human.