“Anne Fehres and Luke Conroy & AI4Media / Better Images of AI / Data is a Mirror of Us / CC-BY 4.0”

In the white paper, the AI4Media Centre of Excellence researchers (from FIAT/IFTA members Netherlands Institute for Sound & Vision and VRT) take stock of existing practices in European media organisations and suggest possible paths forward, outlining essential elements for a robust organisational AI strategy. It synthesises the outcomes of consultation sessions with professionals from 16 media and technology organisations. All of them generously shared their experience of tinkering with in-house AI development, deploying off-the-shelf solutions, managing user expectations and dealing with skill gaps. One of the key takeaways is that media organisations need to accept and manage a number of paradoxes that come with AI integration:

1. Fast individual experimentation versus slow organisational response – while a few risk-averse employees might be leading the way with fast-paced individual experimentation, the organisation as a whole often moves at a much slower and more cautious pace;
2. Pockets of concentrated knowledge versus lack of organisation-wide understanding – knowledge about AI is not evenly diffused across organisations which might in particular impact organisational conversations about policy, long-term strategy and legislations;
3. Managing resistance versus fear of missing out – the reluctance and fear to work with AI coexist with anxiety about missing out on business opportunities;
4. Need for specific solutions versus resistance to give away too much – the need for customised AI solutions tailored to specific content types comes with inherent uncertainty and risks surrounding data sharing;
5. High-quality expectations versus encouragement of criticality – users often expect sleek and seamless interfaces but disarming criticality cannot become the goal.

The white paper proposes strategies on how to approach these paradoxes, including setting up task forces that aid with knowledge diffusion, formalising engagement with AI through a charter and investing in long-term partnerships based on shared values. The paper also outlines elements of an organisational AI strategy that can help to crystallise the best approach for managing these paradoxes. There is no one-size-fits-all solution for this – depending on their context, values and resources, each organisation will require a different set of rules and principles. To address this, the white paper outlines eight key elements that could form the foundation of an organisational AI strategy. Media organisations can use them as guiding questions or chapter titles to start the conversation.