• In the first of our series of interviews with contributors to our forthcoming news report on trusted journalism in the age of generative AI, we talk to Lucy Kueng, Strategic Advisor/Senior Research Fellow at the Reuters Institute
• In a wide-ranging interview, Lucy talks about artificial intelligence and the changes, opportunities and challenges ahead for the media industry – and its capacity to respond 
• Lead author and interviewer is Dr Alexandra Borchardt.
• The EBU News Report 2024 will be available to download from June

Is generative AI a game-changer for journalism and the media?

Yes, but not just for journalism and the media. It is such a profound shift; the media industry will be caught up in the changes. The media’s reaction to generative AI now bears the scars from the painful transition to digital. It took the industry a long time to understand what was shifting, what that meant, and then which changes to make. And it has emerged from that transition in a weaker position. There has been an explosion of innovation, and profound changes in business models, but overall, the sector has emerged structurally challenged in a way it wasn’t before the advent of digital technologies.

In your recently published book “Strategic Management in the Media” you wrote “the fruits of two decades of painful disruption are being themselves disrupted”. Your confidence in the media’s ability to change seems to be limited  

I think the media has exhibited an extraordinary ability to change – it has transformed itself radically. I don’t doubt the media sector’s ability to change – it’s more that it is extremely unclear at this point what the intelligent response is. We want to act, to equip ourselves to deal with what’s coming, what that looks like is less obvious. If we layer on deep concerns about AI’s impact, plus an AI fatigue arising from 18 months of worrying then we get to a place of ‘threat rigidity’. 

When we are stressed and concerned it’s hard to be innovative, to be objective. I see it at conferences, with consulting assignments: People want predictions, they are anxious and want their anxiety reduced. But this is very difficult because we are at such an early stage in such a big transformation. We are at the dial-up modem stage of the internet. Things will change fast.

Do you think the industry is better prepared this time?

Mentally yes. Definitely. But the scope and profundity of these changes mean responses are hard to design. We can’t wait for something to happen, for things to become clear, but this is a new foundational general-purpose technology.  It’s a platform shift, a new layer that products and businesses are built on. The last shift for the media emerged from the coming together of mobile devices, social media, and the cloud. Generative AI is an amazing tool. It will get cheaper, easier to use, and applied more and more broadly, and other tech advances will get bundled with it. This is all to come. 

How can organizations prepare for these big unknowns?

We know that, as Mustafa Suleyman put it, ‘generative AI is ‘climbing the ladder of cognitive abilities’ fast.  That makes it a ‘discontinuous tech shift’ – it can sweep away the value of some existing competencies and knowledge’.  So, some labour will be displaced. But new tasks will emerge – particularly around bridging and blending synthetic and human input. At one step remove however it is also what the late Harvard professor Clayton Christensen called a “disruptive technology”- it allows new players to enter the market with new products produced in different ways and different value propositions. To respond to this threat media organizations must become more agile, and open to rethinking what they offer and how they produce it.  
 
Are you worried or excited about this development? 

From the perspective of a researcher, it is fascinating. I can’t wait until I have a generative AI personal assistant ... But I am concerned.  For the media business, it’s a very uncertain time, Competition is already fierce, and this brings in new actors which will increase competition further.  A flood of synthetic content can worsen the decline in trust that quality media are already struggling with. And for commercial media, more products mean less opportunity to raise prices and ARPU (average revenue per user).  

So, you think media organizations should slow down a bit?

No, the issue is not to panic. But critical also is to do the right things at different levels of the company. At the top it’s about setting strategy. Looking at developments in a broad sense, at what matters most in your organization, and identifying which activities lend themselves to generative AI, where is the biggest potential leverage. On top of that, there’s setting policy: developing guidelines, codes of practice, and putting guardrails in place – where would the application of Gen AI tools create risk?  Then there’s protecting the value of your content assets. Large language models need two things, computing horsepower, and quality content. Media organizations have that content, so have an asset that LLMs need. They have leverage and need to ensure they use it. It would be great to see more collaboration.

You were talking about what needs to happen at the top. What needs to happen at the middle level?

As Charlie Munger said, ‘just learn, learn, learn all the time’ - one of the positive impacts of generative AI is that it has put organisational learning at the top of the strategic agenda. We need to experiment. By using these tools, we build up knowledge. So, in the middle it is about building the expertise, finding efficiencies and testing use cases, letting knowledge bubble up from the periphery. Start small. Identify the things that lend themselves to this technology and getting people to try the tools and find the places it can move the needle.  

That’s why everyone is appointing AI directors now. Just to shed responsibility at the top?

I think it’s more a case of orchestrating the response, ensuring all bases are covered and mitigating risk. With digital, we had digital directors for a while too, but eventually those roles disappeared: every role had a digital component.  

Content creation is at the very heart of media companies. Could this make them react more slowly than other industries, just as they were too slow to react to classified advertising migrating to platforms because it was at the heart of the publishing business model?

The media industry is well-set up for a speedy response. Generative AI has huge potential for sectors that do ‘knowledge work’ and have well-structured data sets. Both are true for the media. Plus, they have learned from digital it is better to move fast than wait to see what emerges. In this respect I am optimistic. 

Do you see any red lines for media organizations?

It needs to be made super clear internally, when generative AI is used, what it is used for, how it is flagged to audiences. It needs to be very intentional. You must be extremely careful what data you load up to Chat GPT.  Also important is to know the strategically critical areas where there could be reputational damage. And, of course, be smart in dealings with the tech majors. The media industry gave away the keys to the kingdom once. That shouldn’t happen again.

Will public service media be better equipped to deal with the challenges? They have the advantages of size and trust. 

Public service media have an amazing strategic advantage because their brands are so well known. They stand for quality and are deeply anchored in the communities they serve. But they are constrained. Public service media are large complex organizations under high levels of scrutiny with relatively constrained strategic options – it’s harder for them to roll with this, to try out the new technologies and see what happens.  Smaller players can simply give things a go.