Three predictions for the future of media distribution in the age of AI

At the same time, I have been extensively experimenting with AI over the past few months. Tools like Pinokio, LM Studio, and Jan make it possible to run AI models on a home computer without requiring deep mathematical expertise. These models already demonstrate impressive capabilities, and we know this is only the beginning. At the same time, we are witnessing the rise of AI-enabled hardware in consumer devices, such as Apple Intelligence and Microsoft’s Copilot+ PCs, alongside chip manufacturers implementing highly scalable computing architectures with specialized AI cores.

When we consider these two developments – the increasing availability of AI models and the rise of high-performance consumer devices – it becomes evident that AI will have a profound impact on media consumption and distribution.

AI in media

Media and technology are inseparably linked – technology enables media, but it also imposes limitations. First comes technological innovation, followed by applications, and finally, demand. Innovative media creators and journalists harness new possibilities creatively, while business leaders establish the economic structures that sustain this ecosystem.

Over the past 100 years, the biggest game-changers have been electronics, integrated circuits, computers, and, most significantly, the internet. Each of these innovations has fundamentally transformed the media landscape, paving the way for new forms of communication, production, and distribution. While we speak today about “New Media”, the next stage will be “Synthetic Media”, generated entirely by AI and indistinguishable from human-made content.

The rapid progress of machine learning is driving an unprecedented technological shift. In the United States, a US$ 500 billion fund has been allocated to expand machine-learning infrastructure – an amount nearly half the country’s annual education budget. Meanwhile, the European Commission is focusing on regulation with the EU AI Act.

At present, AI primarily plays a supporting role in media production. Transcription tools assist newsrooms in their workflows, and automatically generated subtitles have reached a quality level that is now acceptable for users. The next logical step is AI-generated avatar signers. Live encoders optimize video using machine learning. The first implementations of LLMs (large language models) for generating weather and sports news are already in place. On YouTube, where quality standards are often lower, fully automated video content has become ubiquitous.

Three theses

AI is ushering in the next stage of digital transformation – the post-internet era. The long-term impact of this AI revolution remains uncertain but let’s try to think beyond the status quo. I will outline three key developments in the following theses.

Thesis 1: LLMs will become personalized gatekeepers and navigators for information and media

More and more people are already replacing traditional search engines with tools based on generative AI. The eloquent responses generated by these AI systems inspire a high level of trust – whether for recipes, help with maths homework or even medical advice. LLMs are no longer restricted to their training data; they can access real-time information from the web, including news, weather reports, and scientific publications. These models are available as cloud services or can run autonomously on personal computers and smartphones.

The traditional gatekeepers and dominant players of the online world are losing relevance. AI agents, trained on customized corporate LLMs or personal data, navigate vast amounts of information, filtering and structuring it in a way that is tailored to the user. This shift gives individuals more control over their media consumption and reduces reliance on search engines and their algorithmic biases. However, the consequences are profound: business models that rely on controlling access to media and advertising-based revenue streams are at risk of becoming obsolete.

The very nature of media platforms is also set to change. AI agents are already curating and aggregating content. If an efficient system emerges that allows creators to build direct commercial relationships with consumers, platforms could become redundant. If it is already possible to order shoes directly from a production belt in China, then enabling direct distribution of digital media should be even easier.

Thesis 2: Deepfakes will become the default – generative models as the new interface for users

This transformation can be divided into two stages. The first will see a flood of AI-generated content. Users will be overwhelmed by AI-generated media. This includes augmented content like automated weather reports, fully synthesized productions, and even manipulative ‘fake news’. Generative AI will increasingly take over the role of traditional media producers, creating content at unprecedented speed and scale.

In the second stage we move to full personalization. Media will adapt dynamically to individual users. The length, tone, and focus of news videos will be adjusted in real time. A weather forecast video will be generated on demand, with the voice, appearance of the presenter, and background customizable to the viewer’s preferences. AI models for image, audio, and video manipulation will become even more advanced and widely accessible. We already have AI models that can write scripts, compose music, and generate entire films. How long will it be before exciting movies that are tailored to match your precise personal tastes can be created on demand? The personal assistant device will act as a creative machine, capable of manipulating and generating high-quality, personalized media instantly.

As the boundary between real and synthetic content becomes increasingly blurred, the question of authenticity will be more critical than ever. Who created the content? When and with what intent? Media providers, platforms, and users alike will need to establish reliable mechanisms for verifying and labelling content. Without such measures, we risk a media landscape where truth and deception are indistinguishable.

Thesis 3: Broadcasters will lose their dominant position and must reinvent themselves as certified content brokers

Today, broadcasters plan, research, produce, and curate content for distribution. However, this traditional model is being disrupted by AI’s ability to generate and personalize media. In a future where AI takes over content customization, the entire media production and distribution process will be radically redefined.

The role of media providers will shift towards curation, contextualization, and verification rather than pure production and distribution, which will be increasingly automated. Traditional media houses may evolve into brokers, acting less as content creators and more as mediators and analysts. They will no longer own the distribution but instead serve as a trusted entity between algorithms and audiences, leveraging their expertise and brand identity to maintain their relevance.

In this landscape, small, specialized content creators will also have an opportunity to thrive. They will be able to produce high-quality, niche content and deliver it directly to their audiences, bypassing traditional media gatekeepers.

The once dominant role of broadcasters, established during the electronic media revolution, is already being challenged by platforms. In the long run, broadcasters will lose their prominent status entirely. One opportunity for survival is to transform into media brokers, leveraging their editorial expertise and actively pitching their content as valuable assets for AI-driven ecosystems.

The human factor

Am I taking this argument too far? Am I underestimating the human preference for analogue experiences, for handmade content, just as people still cherish vinyl records and face-to-face interactions? Or, conversely, am I being too short-sighted, ignoring the possibility that a future super-intelligent AI species will have already internalized all conceivable films, books, and music as part of its baseline training?

I welcome any discussion on how the publication and distribution of media will evolve in the short, medium, and long term.

This article first appeared in Issue 65 of DVB Scene magazine.

---

Remo Vogel in addition to chairing the DVB Project, is responsible for the strategic development of distribution technology for Rundfunk Berlin Brandenburg, part of the ARD network in Germany. His focus is on hybrid systems for programme publication. He also leads DVB-I-related activities for both the EBU and the Deutsche TV-Plattform.