Generative AI in the Media [VIDEO]
Lessons from Lucy Kueng for every industry
In part 2 of my conversation with Lucy Kueng, we dig into Generative AI in the media industry — why it’s on the cutting edge, and what every other sector can learn from it.
The full 25 minutes is really worth your time. But let’s start with a few statistics Lucy shared from the Reuters Institute at Oxford about how Generative AI is being used in the newsroom today:
Key Takeaways
In the rest our conversation, Lucy and I look at specific case studies of different firms taking different approaches to AI.
Key highlights of the video:
0:00 Why these are still early days for Generative AI
1:43 How newsrooms are using Gen AI (the latest research from Oxford)
4:00 The job reporters hate that AI is now doing
6:12 The challenge of protecting intellectual property from LLMs
7:28 Axel Springer cuts a deal with OpenAI, while NY Times sues: “I suspect they’re giving away the crown jewels too cheaply.”
10:08 Great leaders: “high thinking, low ego”
14:22 Debating the lessons of CNN+
19:45 How CNN’s new CEO is putting digital at the core of CNN
21:20 What the gradual death of print tells us about media disruption
23:52 Lucy’s book recommendation
Full Video (Part 2)
Watch the whole 2nd half of our conversation below (or scroll down to read the transcript) —
Catch Up on PART 1! (if you missed it last week)
TRANSCRIPT (part 2)
David Rogers (00:00):
Alright, we could keep going on this, but let me pivot to a different direction, that of course is on everyone's mind these days. It's the question we all ask and that we also all get tired of hearing about. It’s Generative AI. But I do want to actually bring it up with you, Lucy, because I think actually the media industry is one of the sectors where I think it's most interesting to look at this latest technological wave at this point in time. “These are still very early days” is what I would sort of preface by saying, for Generative AI. I think we can learn a lot from the last decade of what happened has happened with predictive artificial intelligence and how impactful it's been on the economy, but how little we knew in the early first couple years of where that impact was going to come.
David Rogers (00:53):
As a result, I advise business leaders that 2024 should be a year focused on, for most of us, really focused on experimentation and learning around Generative AI. Don't think, “Oh, we're going to unlock this new source of value, this new revenue model, this new huge cost savings or something.” If that happens, wonderful. But organize and focus your efforts right now about experimentation and learning. Would you agree with that? And more to that… Have you seen any examples of great examples that you see companies who are really getting in there and trying a lot of things to really take a “discovery-driven” approach (to borrow a term from my friend Rita McGrath) to Generative AI?
Lucy Kueng (01:43):
The media industry has gone in heavy and hard. I mean, we did research at the institute in January and just to give you some data, 92% are using it for backend automation, 76% for distribution recommendations, 72% percent for content creation with human oversight, 62 coding, 71 identifying stories. So, they've really gone in hard, but there are reasons for that, which you partially alluded to. Firstly, Gen AI has enormous implications if your work involves knowledge, content creation and data, and it's got all of those. So, it's…
David Rogers (02:24):
Creative roles, as I think of it, creative white collar roles. Whether it's photography or copywriting, news writing… and software coding has been one of the earliest sectors. But a lot of that is sort of at the heart of the media industry. So, go ahead.
Lucy Kueng (02:38):
Exactly. So, with this early stage with, and you're right, we're kind of at the stage of with the internet dialup modems, it is very early, but the media's already in the eye of the storm. It is a really clear and present danger for the media industry because it removes entry barriers. Anyone can produce professional looking content, be it text pictures, fake content, whatever. It increases competition, it decreases their ability to raise prices. So, it is a massive danger for them. And they were so burnt, so scarred by the shift to digital, they are determined not to repeat this experience. So, this is why last year was the year of panic and frenzy and actually the last six months it's been kind of incremental innovations. As I said, it's seeped into everything.
David Rogers (03:30):
What would, any examples you've seen of a media business or a publisher who's doing some really interesting things or testing some interesting questions about, “Hey, could we use this here or in that part of our work or really think differently about our business?”
Lucy Kueng (03:48):
It's really being used, David. I mean that's the interesting thing. So, if you take VG, the company you know so well,
David Rogers (03:55):
A Scandinavian news publisher for many years.
Lucy Kueng (04:00):
Yeah. So, they found with digital that news readers love little summaries at the top of articles saying what's going to be in it. They also know journalists hate writing them, hate writing, but AI does it perfectly. So, AI does all of those and has been doing that for about six months, eight months. And essentially they put a note at the bottom of the summary, that it was written by AI. And the issue for them is when do they have to stop putting the footnote in? When is that automatic?
David Rogers (04:31):
Yeah, when is that assumed?
Lucy Kueng (04:33):
So, that's absolutely standard practice. I mean in a lot of news organizations, I think what's really interesting is the news industry actually because of its transformation, it's very well set up for this now. So, they have very clear digital workflows and they also, everything is data informed. So, they've got this amazing basis now and they're also, they've developed quite a strong competence in data journalism. So, some of the interesting work I've been doing is they've been building tools so that newsrooms can work with big data sets. So, it can go through enormous census material or criminal material and identify stories. It's a bit like finding new drugs. Technically it could have been done before, but it took a handful of experts years. Now it can happen in a minute.
David Rogers (05:19):
Right. Now machine learning can dramatically accelerate it.
Lucy Kueng (05:24):
So, something… a bot that can search official statistics. Yeah, I think, and what's really interesting is if you check out channels on YouTube, like News GPT, I dunno if you know this, is a 24-hour completely AI, Gen AI…
David Rogers (05:41):
I've seen one of those. I forget, it was a different brand name, but yeah, 24 hour for the news. Yeah.
Lucy Kueng (05:46):
So, it has gone in, it is being used, but what we're seeing is quite interesting strategies around it and it is being striated according to layer in the organization. So, the challenge at the top is around two things really, policy and guidelines. So, absolute clarity, when do we use AI? When do we not use AI? When are we transparent about it, right?
David Rogers (06:11):
Right. How do we disclose it?
Lucy Kueng (06:12):
For example, you can't put an unwritten article into chat GPT to edit it because then that data, which has probably not been through legal protection, is then in the LLM forever, it's accessible. So, that all kinds of safeguards…
David Rogers (06:28):
Right? What goes into the public LLM versus… unless you're building your own instance. That's a huge enterprise issue, that as an individual you don't necessarily think about.
Lucy Kueng (06:38):
What you around, what you also see is like Reuters news agency or other big organizations, they've actually set up their own LLMs either with Google OpenAI… So they've got their own, so that risk has gone. So, that's one issue. The bigger challenge is actually protecting IP. This is a massive, massive issue. And so, you think this is an industry that was completely rolled over by digital disruption and massively weakened by the platforms, by the tech majors. And they are aware there's a big risk of this happening. Again, because if you're trying to build an LLM or Generative AI system, you need two things, massive computing power and very high quality data content. And that's one thing the media industry has. So, for once they have
David Rogers (07:28):
On that very point and kind of back to what's our role? Do we hold the jewels for ourselves? Do we distribute through the new tech player so to speak? Any thoughts on, I'm sure you know the details of this deal more than me, but I believe it's Axel Springer is one of the bigger traditional media businesses that's reportedly struck a licensing deal to say, Hey, we are going to legally allow you to, I think it's with OpenAI, but a Generative AI company. Yes, you have our permission to use our news articles to train and yeah, you can create AI news reports based on our own reporting. We're getting a fee for that. What do you think about that strategy?
Lucy Kueng (08:13):
I think it's a terrible shame, because I mean it's essentially essentially divide and conquer, right? Someone does a deal with, someone does a deal with OpenAI, they don't know what prices the other people are getting the value of this information. I mean they cannot build these LLMs without this. And if you look at the New York Times…
David Rogers (08:30):
And they don't want to hire a thousand journalists to churn this information out every day. That's not the preferred business model of a Google or an OpenAI. So,
Lucy Kueng (08:38):
They couldn't. And if they train those models on synthetic data, the output's going to be rubbish. So, this is gold dust for them. And if you look at the lawsuit, the New York Times lawsuit against OpenAI, what you saw is they did a common crawl, that's how they got the data. But they showed the weightings and the weightings given to the New York Times content was much higher than for anyone else. So, so what my sense is, I don't think they will, but I think it would be an amazing point at which the media industry could come together and say, this is the value. This is of our content. You are very happy to have it. You can't have it free. We're not opening the doors to the kingdom as we did last time around. But instead for resource constrained organizations, 10 million, 50 million, it's very attractive. But I suspect they're giving away the crown jewels too cheap. So
David Rogers (09:37):
Speaking of these kind of big leadership questions, let me kind of pull back a little from highly dynamic, and again, Generative AI in the media industry adds in on top of all the disruptive forces in other industries. All these questions of intellectual property and copyright and things that I believe are going to likely need to significant rewriting of the legal underpinnings of intellectual property law, but so fascinating to see where this goes next.
Leadership, though. We've talked at various points here about these difficult decisions leaders have to make as they're thinking about strategic decisions, who to partner with, what things to do yourself at what timing, at what point do you bring in new talent? At what point do you change the organizational model? How do you respond to shifts in competitive dynamics? A lot of leadership of course is about, and I would say increasingly as organizations become more fluid and less sort of top-down, decision making leadership as smart leaders have told me is less these days about making all the decisions, making the most important decisions.
David Rogers (10:46):
And it's more about constantly communicating where we're going, what we're trying to achieve, empowering, pushing, decision making down as you can. So, as we think about that, communicating a side of leadership, how leaders lead others, what do you see in the most effective leaders who you've seen who have really empowered, motivated, inspired and gotten the people in their organizations to drive change and also have been able to bring along and kind of win the support hold on, the support of if it's investors, if it's a public company, other stakeholders on the outside, you also have to hold their support while you're trying these really difficult things to pursue ongoing transformation in your business.
Lucy Kueng (11:31):
Good question. I'm trying to frame this from your audience, which is a diverse set of sectors, geographies. The leadership positioning for top leaders is essentially this combination of high thinking, low ego. So, you need to have a high IQ because this is really complicated stuff to navigate, to understand, to work out what the hell we're going to do. I mean find strategic clarity in such an uncertain and choppy world. So, high IQ. But what's interesting is that's become kind of table stakes. You don't get in the game without it, but you don't really succeed as a leader. It's the EQ that matters now. And this is being actually pretty evolved as a person. This is quite interesting. You need to be very mature because
David Rogers (12:25):
Which is historically when we think of media tycoons, at least EQ and emotional maturity is not what maybe comes to mind,
Lucy Kueng (12:35):
But what's completely consistent. I mean in the research I've done for the last three books, all the leaders who are doing this, well share one comment in slightly different phrase, but essentially the real breakthrough is when I admitted to people how hard I was finding this myself and I needed a coach or I need to go and learn something because that really encourages the next layer to modify up, it's okay, it's okay to admit this is really tough, complicated stuff. It's tough for me that some of the most important knowledge is now in the youngest heads in the organization. I have to listen to them, they have to explain it to me. But the minute they gave into that and were open about it, that was when the real transformation started to take off. So, it's a kind of different variant of this authenticity. And David, the words that came out of these absolute alpha individuals was like, I had to be accessible. I had to be vulnerable. Very, very interesting. And actually it wasn't that they read a book, it's just they're smart individuals. It dawn, and in terms of maturity, it was about I understand my own limits here and I know when I need other people and I know when I have to be honest about that. Fascinating, really,
David Rogers (13:54):
That's so fascinating. One of the words I've come back to frequently looking at effective leaders today is humility. And that plays out in different ways. And part of that is sort of knowing the limits of what and how smart you are and what you're going to be able to figure out. But I think it's also very much about your interactions with others as you're explaining. So, well, let me pull up one more case example that it maybe is to me a story of a lack of that eq, a lack of that humility. Maybe a warning sign was from CNN and the warning sign would be from, I want to be clear what happened recently and I'm kind of interested in what might come next for CNN, but I've written in my book and in a recent issue of my David Rogers on Digital about CNN+ as an example of a very specific kind of failure, which I see in other industries as well of top leadership coming up with an idea, actually identifying an important strategic kind of problem to solve or an opportunity to consider.
David Rogers (15:05):
And then getting wrapped up in this traditional planning and decision-making process of like, okay, what do we do? And coming up with a very specific idea that's okay, this is the big solution. And the very top people have sort decided this is what the product's going to look like. This is the strategist, this is what it's going to do and how it's going to play out in the market and become sort of sufficiently convinced of their own self-confident that everything just gets locked in and we just barrel ahead. And in that case spent $300 million just preparing for launch of that streaming product and really not at all following the principles of testing, validating, experimenting, building iteratively, et cetera, which I talk about rapid experimentation and how do you build that into organizations. So, that turned out to be a very failure. It was a very poor start, dramatically a weak start to the product.
David Rogers (16:01):
And then that time with the change of leadership, they decided it wasn't even worth continuing to plow hundreds of millions into it and they shut it down. I'm interested to see now what's happening next and it's interesting to see with the new leadership, I think some sort of signs that, believe it or not, they might be learning, they may be taking a different approach, trying some things small first, but what do you see about, I mean that is a brand I think that has some weights and resonance and meaning to people. It's a brand in the space. I would hope that they could find a strategy, a path forward, but what do you see or what might you recommend for someone like that, a legacy brand trying to stay relevant in the era of cord cutting, in this case a news brand, not a Disney or an HBO, but a journalism streaming.
Lucy Kueng (16:58):
Yeah, I mean I think I'm kinder than you are. The CNN+ story, I mean this was really Bergulman's valley of death on steroids. Because the unique nature of the streaming model meant they had a really narrow strategic channel that was available to them. And on top of that, they had a new leadership coming in who by definition were not going to, there was a big leadership fight. So, the new people coming in were not going to like what the people before were doing. So, the people before were rushing against the clock to launch it and give it time before it got dismantled. So, a more slow evolution approach, they didn't have the scope of that. So, I think the problem you have there, which is kind of essentially the problem the newspapers had, but much more extreme is that 50% of their revenues come from those subscription either must carry fees from the cable stations, but the minute they put them on a putative new DTC subscription product, the risk is 50% of their revenues just disappear overnight. And with those will go the advertising revenues. So, what they were trying to do as a straddle strategy, not disrupt the core market that was bringing in all the revenues and then build up a new one gradually. And what you saw they then did is this kind of adjacent product, a softer product around the edges?
David Rogers (18:32):
Yeah, I mean they stayed within penciled within the lines of the legal contracts. They didn't try to break or exit their contract. They said we're going to create a totally new set of content. They weren't going to repurpose any of the content, which is part of why the expense was so high. But there were a lot of business model assumptions baked in and unless you knew you had years of runway, which they obviously knew they didn't. As you said, I think that very much I agree that's part of why they rushed and sort of went so headlong. But I mean looking forward, do you think if they've got, Mark Thompson of course has come in as the new leader of CNN and he's, he's done great work with other legacy news businesses, New York Times and BBC, most notably in the digital era. Would you say, gee, I hope he just gives another run at something very similar to CNN+, would you say? I think they should reconceive the market a little bit differently. If you could give them one word of advice to fight for a certain approach.
Lucy Kueng (19:42):
I think Mark is a brilliant, whichever executive search person came up with Mark amazing because I mean if you look where he did, he did this at the BBC, which has to be one of the most complex organizations in terms of stakeholders. Stakeholders hate each other.
Lucy Kueng (19:58):
Unbelievable. Then he did the same at the New York Times so he can deal with complexity and intransigent bits of bits of the jigsaw to move around. What he's doing is essentially he's doing the classic, he's maintaining the old, but he's building out the new, you can see that happening. He's putting digital at the core, but so he's integrating domestic, international, and digital and he's investing infusing product and data everywhere and he's taking costs out. So, it's a kind of gradual strategy. Firstly, digital is becoming the center, the other parts are being shrunk into that. But it is really interesting what you can see all in the media industry, all of these transformation strategies break down into three levels, which I always build a transformation stack. The bit at the top is building out new businesses for the new world. The bit in the middle is holding onto the old businesses to finance that as long as you can, and the bottom bit is kind of building all the competencies and cultures you need and he's kind of what he is doing. And there was a really interesting quote where I think this month he said the TV lineups are the jewels in our crown. So, he, he's still, and a long time ago for the last book, I did an interview with him and he said the print print is still keeping the lights on. We can't turn the lights off yet.
David Rogers (21:28):
This is for the New York Times for because BBC doesn't have print. Yeah, yeah, the print edition, there's a reason why they still have it. It's not their old fashioned, it's just economic sense still. It makes
Lucy Kueng (21:39):
Huge economic sense. And that was really very interesting, David, beginning of this year or maybe going back when interest rates suddenly went up, cost of living went up, inflation went up. You suddenly saw how dependent a lot of the digital darlings were actually still were. They were still dependent on the old print products because newsprint went up so high and shipping those newspapers around costs so much. And suddenly the financial figures went down
David Rogers (22:06):
That calculation starts to shift, and …
Lucy Kueng (22:10):
Print became unviable. And then it was clear digital wasn't yet strong enough. So, I think it's the same playbook, but it's a bit slower. But I do think CNN+… there were added dimensions of “Succession”-type executive drama.
David Rogers (22:33):
I think the “Succession,” sort of pushed them to say maybe if they had the instinct to say, “We can't afford the time to sort of build product market fit. We've just got to dream up the best possible product and just hope we can land the plane.” And it was ugly.
This has been great. I want to wrap up with just a few last, as I told you before, speed round questions. Just to share some parting thoughts and wisdom and suggestions of your own for folks in our viewership. So, four questions. If you can answer, complete the following:
First question, “The most fun part of the work I get to do is…”
Lucy Kueng (23:25):
It is working with strategic leaders going through this. I mean, it is such a privilege and use the word humbling. It is so much harder to do this than right about it. I'm sure this is absolutely extraordinary. I mean the complexity, but also when you add value, when things get better, it is an amazing feeling. So, yeah, that's that. Perfect.
David Rogers (23:47):
So, fun. Alright, I couldn't agree more.
“One book that you've enjoyed recently and would like to recommend is…”
Lucy Kueng (23:56):
I would read The Coming Wave, which is by Mustafa Suleyman, who was the founder of I think DeepMind, which was bought by Google. He is now. And it shows not only where AI is taking us, but where AI combined with synthetic biology, combined with nanotechnology. Quantum is taking us, I don't understand all of it, but it's helping me understand the landscape.
David Rogers (24:20):
Yeah, yeah. Excellent. Great, great, great suggestion.
“Something I have changed my mind about in a significant way is…”
Lucy Kueng (24:30):
I can't answer this. My problem is I'm a chameleon, so I'm a Brit who lives in Switzerland. I work a lot in Norway and in Germany and here. And one of the problems if you live abroad is you begin to understand every possible point of view. So, my challenge is the other way and kind of working out: what is my non-negotiable core beliefs?
David Rogers (24:49):
What is one thing you haven't changed your mind about in the last week?
Lucy Kueng (24:53):
I can see pros and cons everywhere.
David Rogers (24:56):
Okay. Okay. That's fair. We'll, strike that question.
“One piece of advice that I would give to my younger self is…” if you could go back in time.
Lucy Kueng (25:06):
I think, I don't mean danger, but: Take bigger risks. Actually, if I look back, the things I regret is when things were offered to me. And to be honest, I chickened out and always a mistake. And even if you fail, that learning is where you really learn about yourself, where you flourish, where you don't.
David Rogers (25:27):
Great. Great advice. Wonderful. Thank you so much, Lucy. Let me just end by asking you to let folks know where can people find you and learn more about your work, your writing, your speaking, et cetera.
Lucy Kueng (25:43):
Yeah. Well, David, this has been a pleasure. I always enjoy our conversations. This one has been great. Absolutely. Yeah. My website, LucyKueng.com and also on LinkedIn.
David Rogers (26:00):
Wonderful.
Yeah. Lucy posts regularly on LinkedIn and one of the brilliant people who I'm always, I always like her posts both because I like them and also because I want the algorithm to make sure to show me the next one. There's always great things coming up from what she's been doing.
Well, a huge thank you, Lucy. It's a pleasure. Always a pleasure to chat and get to hang out, but particularly fun in this case because we get to share it with all of my subscribers who I know are going to get so much benefit and joy I think out of this video. Thank you so much for an amazing conversation, and thank you. Please keep in touch.
Lucy Kueng (26:39):
I always enjoy your expertise. Take care.
David Rogers (26:42):
Bye. Thanks, Lucy.
NEW BOOK:
“THE DIGITAL TRANSFORMATION ROADMAP: Rebuild Your Organization for Continuous Change”
ORDER NOW:
Hardcover: https://amzn.to/41U85dl
Kindle: https://amzn.to/3OWD437
Audiobook: https://bit.ly/DXR-Audiobook
Bulk orders up to 60% off: https://bit.ly/DXR-bulk-orders