Â鶹Éç

We are at a tipping point in the debate about how far public service media organisations should embrace personalisation. In an already fiercely competitive global media landscape, public service media organisations are under pressure to offer audiences ever greater levels of personalisation, increasing access to and discovery of media content. After a century of the 'one-to-many' model of broadcast delivery, digital technology is now shifting towards hyper-personalised and interactive media. What audiences see is increasingly driven by algorithms, while generative AI can create or adapt content to individual inputs in real time.

Critics have argued that recommender systems and customisable media formats undermine core public media values including universality and diversity, diluting its ability to promote a shared, common culture. When algorithms decide what’s next, the question becomes, what are they optimising for?

So, how should public service media adapt in this rapidly shifting market, whilst staying true to its core public service mission?

Addressing the tension around personalisation in public service media

These are not necessarily new questions. The drive towards greater personalisation in media has been apparent for decades. Public service media leaders have talked of alternative approaches such as â€˜public service algorithms’ for many years.

Yet, the outcomes of such alternative approaches remain elusive. A recent report found that personalised recommendations on UK public service broadcasters’ streaming platforms were , such as arts, international content, and religion.

The wider technological shift is also becoming more acute, with the rise of AI agents predicted to bring a new wave of hyper-personalisation.

As technology and audience expectations evolve, public service media face a central dilemma: how to advance personalisation in ways that still meaningfully deliver on their public service mission.

Is personalisation and public interest a false choice? What we’ve found

Public service media organisations, particularly in the UK, have tended to champion the value of personalisation while academic researchers have been more cautious.

Academics have argued that personalisation is rooted in serving individual preferences, in contrast to public service media ideals, which are traditionally built around collective notions of the citizen or community. Others have highlighted risks over the loss of shared cultural experiences, reduced exposure to diverse ideas and viewpoints, as well as concerns about data privacy and algorithmic bias.

Public service media organisations, however, have argued that personalisation is key to helping audiences discover public service content in the digital era, and that it can even be a differentiating USP against commercial competitors. They are therefore actively experimenting with how to use personalisation to serve public service outcomes.

For example:

• German public service media organisation ZDF has incorporated  and novelty into its streaming recommendations. 
• The Â鶹Éç has embedded editorial values into its recommender systems through .
• Swedish Radio has developed an  so audiences can ask questions about the news and get personalised answers sourced only from Swedish Radio’s own journalism.
• The Â鶹Éç is currently developing an AI-powered personal learning companion for Â鶹Éç Bitesize that can support interactive and tailored learning in a public-service driven way.

These examples demonstrate that personalisation and promoting public service values are not binary opposites. They can be mutually reinforcing.

Practitioners and researchers must therefore collaborate on pragmatic paths through the personalisation dilemma, finding more room for nuance and context.

A critical research agenda for personalisation and public service media

Our report outlines  move forward:

• Metrics for public service optimisation: Research to support the development and adoption of measurement frameworks that can optimise recommender systems around key public service media values, rather than commercial metrics such as engagement.
• Scalable editorial safeguards for AI-driven personalisation: Research into how public service media can maintain editorial integrity when using automated systems to deliver personalisation at scale – such as product, educational, or cultural interventions.
• Public service media personalisation by platform intermediaries: Research into how personalised recommendations and curation by third-party platforms affect the availability and discoverability of public service media content.

In summary, we need a deeper analysis of the possibilities of personalisation in the public interest, in both theory and practice. There is a conceptual challenge to reconcile personalisation with core public service values.  We need more insight into practical implementation issues, such as the cultural and organisational infrastructure needed to support positive use cases.  The personalisation dilemma is also not a challenge that public service media can address in isolation. Personalisation is embedded within the broader stuctural power dynamics and platform logic reshaping the media industry. We therefore need research that contextualises personalisation within these platform dynamics.

So, what’s next?

These challenges require urgent attention as personalisation becomes more sophisticated and the competitive pressures on public service media only intensify.

In the next phase of our research, the Responsible Innovation Centre will work with public service media organisations, researchers, and policymakers to address these priorities.  We hope this report can provide pathways forward not just for collaboration between industry and researchers, but also to engage with policymakers, civil society, technology companies and the wider public to make personalisation work for the benefit of everyone.