How meaning gives content its content and what AI contributes to it
is the title of a roundtable podcast around the study “Man or machine – who writes the better content” with study author Prof. Peter Gentsch, content marketing guru Mirko Lange and content intelligence expert Roland Fiege from rellify™ on 10/17-22.
Here you can find the transcript of the one-hour conversation: part 1, part 2, part 3
Part 1: New wine in old bottles: What is actually the content of content?
Roland Fiege: Welcome back to the AI in Search Podcast with interesting topics around AI and content marketing. I welcome that with me today on the podcast Professor Dr. Peter Gentsch and Mirko Lange, CEO, founder of Scompler, a communication management platform that is great for managing content on and for the web. Please introduce yourselves briefly in one or two sentences!
Peter Gentsch: I’d be happy to. For more than 20 years, I have enjoyed leading a double life between theory and practice, between my Chair of Digital Transformation and International Business and in real life for companies in the field of data, analytics, AI, for which I try to bridge the gap from theory to practice.
That’s sometimes not so easy, but we’ll certainly talk more about that today.
Roland Fiege: And then we have Mirko Lange from Scompler with us. Tell us in one or two sentences what you do with Scompler.
Mirko Lange: Well, I actually come from strategic communications consulting. And in that context, I’ve always suffered from the fact that, of course, we’ve created the most amazing content and communication strategies, 500 page power points, but then somehow found that companies are not able to put this into practice. That’s when my motto was born, “It’s not enough not to have a content strategy. You also have to be incapable of implementing it.”
So we then started to build a small software in a big project with Deutsche Bahn. And we hit the mark with that. In the meantime, we have found large investors, have more than 300 customers, including six DAX companies. Our largest customers work with 500 people worldwide on our platform and coordinate the entire topic management there and actually the entire process from content strategy to topic identification, topic planning, production planning, and production to analysis.
And because we are, of course, always on boarding our customers on the platform, we have had a great deal of insight into many different companies and how to map content strategies completely and then ensure that these strategies are operationalized in day-to-day business.
Roland Fiege: That means you deal with content curators, with authors, on a daily basis. In other words, with the people who are responsible for topics and the texts that then appear on corporate websites. That’s why we invited you today, because we want to talk about Peter’s study “Man versus machine – who writes the better content?”. Because machines are already capable of writing meaningful texts today, which will probably bring a huge change to your customers as well. Or at least raises the questions, can you save costs, can you write the better content? Content that is better understood and evaluated by people and also better by search engines? That’s what we want to discuss today.
Peter – perhaps you could briefly present your study, what exactly it was about and what the key messages of the study are?
Peter Gentsch: Yes, with pleasure. First of all, it has to be said that the man-machine discussion is not particularly new. But there is a new generation of AI, so-called transformer models or foundation models, which at least claim to revolutionize AI.
Classic robo-journalism, in the sense of I can somehow do a weather report or a sports report on a soccer match, has been around for a long time. Or texts on the development of share prices. That’s almost a commodity today.
But the question was, can you also write creative texts with AI? Can you write blog posts? Can you write landing pages? Can you write social media posts? So our approach was to use the new generation of algorithms to see, for once, how easy it is to create content. And, more importantly, what is the quality of that content? How is such content actually perceived by the audience?
In addition, the study should also be scientific. There is a KPI, the so-called Flesh Index, which measures the readability of content. So we looked at various brands, for example L’Oreal, Starbucks, BMW, Telekom, and looked at their original content. So landing pages, blog posts, social media posts. Then we asked the AI to create the corresponding content.
There are two approaches to this. First, I can use an AI to analyze what exciting topics are, for example, in the area of coffee, sustainability, Starbucks. On the other hand, I can tell the tool very precisely what it should be about. This feeding of the tool is the exciting part, it’s called “prompting”. In the past, you had to program complex AI algorithms to do that. Today, it’s become very simple. So we did that, and exciting results came out. We presented these to 100 consumers (without saying this was an AI test), and told them we had two contributions here from Starbucks: Which one do you like better?
Across all industries, the AI-generated texts were found to be at least par or better! And what was so surprising to me, statements like “the text seems more personal” came up. That’s downright counterintuitive that an AI texts “more personally.” And that the flesh index, the readability of the content was higher in all areas than for content created by the agency, by the company, by the editor, whoever. In short, the AI texts read very nicely and were perceived as personalized. But then I showed them to a colleague from BMW, and he said “yes, that reads quite nicely, but it’s not true at all!”. That’s not the current model at all!” And that happens more often, that the AI generates content that is very easy to read, but in some cases is simply incorrect. This is, of course, a problem that we have in AI in many areas: AI can be used very easily and very low-threshold. But of course you have to be careful as hell – who actually ensures the quality of the content?
In any case, it was exciting that this was not just about standardized reports, but about how a brand presents itself empathetically in social media. You wouldn’t expect an AI to be able to do that!
Podcast anhören:
Teilnehmer
Mirko Lange, CEO, Scompler
Mirko Lange, Scompler
“Content Marketing Pope”, “the German Joe Pulizzi”, “Content Strategy Guru”… when Mirko Lange is presented as a speaker at conferences, the laudators find many nice terms.
Mirko Lange is one of the most renowned strategy consultants for content and content marketing in German-speaking countries. He is the founder and managing director of Scompler, a strategic editorial planning tool with more than 20,000 registered users and around 150 corporate customers.
Prof. Dr. Peter Gentsch
Prof. Dr. Peter Gentsch
Prof. Dr. Peter Gentsch is a speaker, entrepreneur and scientist in one person and has been one of the pioneers and top experts in the field of digital transformation, artificial intelligence (AI) and big data since the 1990s. While others only talk about digitization and technology trends, he lives them.
With numerous company start-ups and investments and five successful exits, he is one of the most successful Internet entrepreneurs in Germany.
He not only talks about digital transformation and AI, but shapes and shapes it. He imparts this knowledge and experience as a coach and speaker in companies.
Roland Fiege (Host, Managing Director rellify Deutschland GmbH)
Roland Fiege
managed many international companies as senior director in the field of social strategy and general manager. After his MBA at the Rhein-Neck University in Mannheim, he did research in the area of Social Media Balanced Scoreboard (SMBSC), Social CRM and analysis systems. He taught in the areas of strategy development for social media, monitoring & analysis at various universities. At rellify, Roland brings his experience and passion for marketing, entrepreneurship and technology as General Manager EMEA and leads the rellify team in Europe.
Roland Fiege is a sought-after discussion partner for various news channels and radio editors, such as RTL/ntv, ARD, Deutschlandradio and others. Together with Karl-Heinz Land he publishes the well-known podcast Earth 5.0.
Roland Fiege: Mirko, how do you see it? What is the status quo in content marketing at the moment? We’re talking about texts here, long-form texts, shorter texts, but also social media texts. The media landscape is very, very fragmented. Do people even read longer articles anymore, or do they only read headlines? Where is the trend going right now? Short-form or long-form content? How do you see that?
Mirko Lange: So my standard question to my students is always, “What is actually the content of content?” Then, as a direct continuation, I follow up with the question of whether a whitepaper or a blog post or a long-form is content. And that’s when students say “Yeah right, a whitepaper is content!” and say “No, what’s in the white paper is the content!”
And that’s what I mean by “What is the content of content that we distinguish today?” Let’s take a wine bottle as a comparison, where we design the shape and the color and also the label and then make sure that that wine bottle is on the shelf at the point of sale. This point of sale is the channel or the touchpoint. That’s where the user decides whether he’s going to buy bottle A or bottle B. In a large hit market or in the metro, the bottle is on a 20-meter-long shelf – and most of those standing in front of it may be looking at the price and the look, and maybe even the grape variety. But the range is simply huge and there is a lack of orientation.
But that only ever applies to the first sale. However, we as marketing people, also in product marketing, want people to buy our products again. And now comes the crucial question: “What does that depend on?” On the content, that is, on the wine!
So we have to consider whether we are wine bottle distributors or winemakers. Of course, the wine has to be bottled and could not be sold without a bottle and label. But here we already have a fundamental distinction in response to your question, which I don’t think can be answered in the way you asked it. We have an incredible wealth of applications for content marketing. For me, this term is actually a non-word because we say, well, there’s social media marketing and email marketing and search engine marketing. But how are you going to do all of that without content?
So we can’t see content marketing as another discipline alongside those other disciplines. Obviously, content is a key asset that’s used in all forms of marketing, in corporate communications, in internal communications, and so on. That’s, by the way, also the thinking of Scompler, that we define content as a central set, make ONE central content management, totally channel-independent or discipline-independent. And only then, once we’ve found the story, ask how do we need to tell the story on different channels for different audiences?
Let’s take fast food as an analogy. If you ask me where the trend at trade shows is going, on the one hand it’s going in the direction of convenience, but it’s also going in the direction of commodity. So a burger today is commodity, but it’s also moving more and more toward premium and healthy food on the other side. It is splitting up – and this analogy is 100% transferable to content. And by the way, food is also content that is ingested, “inhaled”. So it’s all about doing the right thing for your target groups, for the topics, for the markets! And we have a huge bandwidth here. And for a small part of this bandwidth, AI is also super suitable!
Roland Fiege: So there are an incredible number of channels and there is content. It first has to be transformed and then reduced, shortened, simplified or prepared for other media and published. What are the main pains that typical marketing departments have these days when it comes to content marketing?
Mirko Lange: So throughout the entire process, at every point they have pain and unfortunately that’s also a bit of our fault. So it starts off with 60% saying they have a content strategy according to surveys. As I said, I’ve seen a lot of that and to my mind, out of that 60%, maybe 10% deserve to be called a content strategy.
Either they have something that’s way too big so it can’t be operationalized or it’s just way too vague. Almost no or very few companies even have a topic architecture that defines what do we talk about, what do we not talk about? Many only proceed on an occasion-driven basis, e.g., they let an AI tell them what we should write about, and the AI then looks at search volumes or search engine rankings, for example. But don’t get me wrong – the data is all helpful, but it also has to be put into context with strategic objectives – how do I want to position myself, how do I want to distinguish myself.
We have therefore developed our own methodology for scoring topics. Of course, our own interests are also important here. A media outlet, for example, works very differently from a corporate communications department. The medium just has to pay attention to the interests of the readers.
We always distinguish between “topic” and “story. The topic is abstract, such as soccer or marketing or sustainability. And the story is then the concrete cause. So how can I heat sustainably despite the energy crisis? We call that a story, and you have to tell it differently for different target groups. There’s often a problem with getting it right, with knowing the target groups correctly.
In production, it’s about creating the text correctly, in different variations, and preparing it to suit the channel. And that’s the least of the pain: in my opinion, this tactical, purely operational activity is the least of our worries. There are well-trained people who know how to prepare it for TikTok or Facebook or Instagram or whatever.
Then when that’s been distrubed, there are issues again in the evaluation. “Reach” and “engagement” are typical KPIs, for example, but in my view they are far from sufficient to judge content.
I always say that what we call content today are in most cases single notes. But we have to look at the melody, i.e. the sum of all publications, which the target groups then also absorb from different media. And communication is more than just output. Now, output usually just means reach or engagement. But we want impact, that is, the next step. So what does that mean? What effect does content have on communication? That’s where I see the biggest strategic or structural problem, that most people stop at output.
This brings us to “snackable content” and the keyword “commodity. That certain types of content are consumed very well, but similar to tissues or panty liners or toilet paper, that at some point it simply no longer matters who the provider is. And then content marketing becomes absurd. Then I have my reach and engagement, but no connection at all to the brand or any business or value-creating effect.
Part 2: Significance, visibility & relevance – opportunities and limits of AI in content marketing.
Roland Fiege: Mirko Lange, you just said very nicely that when it comes to preparing certain blocks of topics for different media, technology is obviously not so necessary because there are many people who can do that very well.
But you also said that there can be added value in finding topics. Peter, how do you see that? Do you see more of a benefit in AI, in the machine shortening and preparation of already finished content, or actually in topic finding, in topic management, that is, at the very beginning of the strategy?
Peter Gentsch: For me, there are clearly two phases. When it comes to finding topics or constructing topics, an AI can only provide support in the sense of augmented intelligence. It is always pretended that AI does something completely different than humans. No, an AI can go out and look globally for certain trends. Like Mirko said, for example, looking for SEO rankings, what topics are coming up, how contexts are changing. But what an AI can’t do, of course, is interpret that data, evaluate it, put it into a brand context.
And of course it’s easier to say, dear AI, look for all the trends and summarize topics from the worldwide web, and then I’ll look at them as a professional and develop a content strategy from them. In the first step, I think it would be presumptuous to say that this can be automated. It’s more about support, a kind of brainstorming that AI can do by intelligently scouring the vastness of the Internet.
In the second step, when it comes to producing content, you have to differentiate, of course: Is this an essay in a great newspaper or is it a small blog post? One thing has to be said clearly: At the moment, production based on AI is possible for relatively manageable content in terms of length. Today, I don’t have to expect my students to produce a master’s thesis, a doctoral dissertation, or a bachelor’s thesis using these new algorithms, because they have the problem of making connections and finding a storyline.
But for individual content modules, for example a blog post for Starbucks, you can use this very efficiently. And that’s why I think the discussion is completely wrong about whether AI can replace content marketing and the editor. It’s more about understanding: How do I actually manage to use the systems intelligently in the various phases with what they can do and what they can’t?
Mirko Lange: In my opinion, we should add another term here, namely “scaling”. So whenever it comes to scaling content, I would recommend using AI, that is, producing content on an assembly line. As the assembly line analogy makes clear, I then have a machine that produces it. But these have to be standard products or products that can be standardized. Because AI is looking for patterns.
So for utility texts, either for search engines or simply to populate a large number of platforms, AI is very well suited. But to make a mark, to make something special, AI is hardly suitable. So I would like to take that thought further as well: If everyone uses the same AI, then content will become totally commoditized – it won’t matter where I read it and from whom!
Peter Gentsch: I would underline two points about scalability: That’s always been the case with stock prices or weather reports. There I have the scalability, even if that is certainly not such exciting content. Or when it comes to product descriptions for e-commerce stores or product reviews. There I have scaling, there I have standardization. I would completely underline that.
But again on the topic of commodity: In fact, the question is, if everyone is using the same AI, these great models that are out there, how do I still differentiate myself? One differentiating factor is, of course, the human being, quite clearly. But I can also feed an AI quite specifically with my domain knowledge, with my brand DNA. And that is the supreme discipline, not just to take an out-of-the-box AI that everyone works with, but to enrich it accordingly.
And as a result, the man-in-the-loop will continue to exist for a while, and that’s probably a good thing.
Roland Fiege: That brings us to the “dangers” or “questions” that arise… If I say I want a text in the area of mobility and sustainability, how likely is it that it will be written very differently for one brand than for another? Probably not, if the machines all use the same data base. Of course, this raises the question of how the brand wants to be perceived out there. And I think, from what I’ve seen so far, we’re not there yet. I think it’s actually very, very helpful in identifying topics, finding topics, and of course scalability.
The danger is that if content is ultimately only produced in order to produce content and otherwise no longer serves any purpose, except perhaps to somehow generate a click for something, then in the end it doesn’t do the brand any good!
Mirko Lange: One example of this is the media system today. It’s often all about producing more and more content, more and more clicks, and then producing more and more clickbait, i.e. headlines that lead to clicks. And that goes into an area where you can’t win anymore. In my view, the industry is completely destroying itself.
As creatives, we have to focus on topic design, on storytelling, on contextualization – and here’s the thing – on enrichment with meaning. To the question I asked earlier, “What is the content of content?”, today I give the answer, “Meaning.”
So we have to explain to people why a certain subject has meaning for them in the current situation. So, if you have no meaning, what are you? Meaningless. And the more content there is, the more the art is in assigning meaning. Assigning meaning goes through connections, networking, creating context. Only when the script is written, so to speak, is it a matter of producing the text. And I hand it over anyway. And to put it bluntly, it doesn’t really matter whether I hand it over to a freelancer, copywriter or the AI. I also believe that the AI can be even better. If the briefing is good, if I can “feed” the AI with the briefing in a really specific way. But for that, there has to be a good briefing first.
Podcast anhören:
Teilnehmer
Mirko Lange, CEO, Scompler
Mirko Lange, Scompler
“Content Marketing Papst”, “der Deutsche Joe Pulizzi”, “Content Strategie Guru”… wenn Mirko Lange auf Konferenzen als Redner vorgestellt wird, dann finden die Laudatoren viele nette Begriffe.
Mirko Lange ist einer der renommiertesten Strategieberater für Content und Content-Marketing im deutschsprachigen Raum. Er ist Gründer und Geschäftsführer von Scompler, einem strategischen Redaktionsplanungstool mit mehr als 20.000 registrierten Usern und rund 150 Unternehmenskunden.
Prof. Dr. Peter Gentsch
Prof. Dr. Peter Gentsch
Prof. Dr. Peter Gentsch ist Speaker, Unternehmer und Wissenschaftler in einer Person und zählt seit den 90er Jahren zu den Pionieren und Top Experten im Bereich Digitale Transformation, Künstliche Intelligenz (KI) und Big Data. Während andere nur über Digitalisierung und Technologie-Trends referieren, lebt er sie.
Mit zahlreichen Unternehmensgründungen und -beteiligungen und fünf erfolgreichen Exits gehört er zu den erfolgreichsten Internet-Unternehmern in Deutschland.
Er redet nicht nur über die digitale Transformation und KI, sondern prägt und gestaltet sie. Dieses Wissen und seine Erfahrung vermittelt er als Coach und Speaker in Unternehmen.
Roland Fiege (Host, Managing Director rellify Deutschland GmbH)
Roland Fiege
leitete viele internationale Unternehmen als Senior Direktor im Bereich Social Strategy und Geschäftsführer. Nach seinem MBA an der Rhein-Necker Universität in Mannheim hat er im Bereich Social Media Balanced Scoreboard (SMBSC), Social CRM und Analyse Systeme geforscht. Er unterrichtete in den Bereichen Strategie-Entwicklung für Social Media, Monitoring & Analysen an verschiedenen Universitäten. Bei rellify bringt Roland seine Erfahrungen und Leidenschaft für Marketing, Unternehmertum und Technologie als General Manager EMEA ein und leitet das rellify-Team in Europa.
Roland Fiege ist gefragter Gesprächspartner verschiedener Nachrichtensender und Rundfunkredaktionen, wie z.B. RTL/ntv, ARD, Deutschlandradio und anderen. Gemeinsam mit Karl-Heinz Land publiziert er den bekannten Podcast Erde 5.0.
Peter Gentsch: Talking about significance is a very ambitious term. I would like to talk about relevance first. And if I now take an AI, what it can do, for example, it simply looks at which content is currently performing well. So very mechanical in the sense of “high Google ranking”. Then it takes those content pieces, mixes them, curates them and comes out with a new article that out-performs the others.
Of course, it’s a highly mechanical process; if you will, it’s recycling. It’s a copy and paste from the AI. But in the end, there is implicitly human knowledge in it, because the people who wrote the contributions have thought about it! And, to put it wickedly, an AI also steals this implicit thought in the articles and takes a kind of “Best of …” and says “look, a new article!”. And for me, that’s one of the things about this new generation of AI: It also imitates creativity to a certain extent!
But how do you see that? Because if we say relevance is what ranks high on Google, then of course that’s a very simple formula. And then an AI can very well generate relevant content, because implicitly the rationality of a human being is already behind it!
Mirko Lange: Relevance is a super nice word, because people always say “you have to create relevant content”. And thankfully you have now defined the reference point for relevance, namely as what ranks highly in Google.
I would now contradict this a bit – and I say this a bit hesitantly – because the algorithms according to which Google evaluates are also very complex today; there is also a lot of consideration behind it. But I think that’s where the key lies, in knowing what’s relevant. And here we come full circle: In my opinion, relevance is always something relative. What is relevant for you, Peter, may not be relevant at all for Roland. So relevance is something individual and specific, target group specific. And today, because this word “relevant” was so intangible for me, I define relevance as: What significance does this have for my life? So what can I specifically do with it now? Does it give me orientation? Does it give me use value or practical value? Does it possibly just give me entertainment or does it give me meaning? So this whole issue of “purpose driven marketing.”
And Google, yes, has a different model of “what” you use something for, so “what do I need this for?” Or “what is this good for,” which is different than “why,” by the way. And, of course, an AI will be able to do that at some point. If we get more specific there, then I also think AI can achieve a very good impact. But just Google ranking would be too shallow for me.
Peter Gentsch: Yes, well, let’s just say “visibility in digital media”. That is, after all, an extremely important currency. And if AI helps me to have high visibility in the digital space, then that may not be high quality and mainstream in some cases and may have potential for optimization in terms of level. But visibility is obviously a currency that at the end of the day probably pays in to conversion, pays in to brand building. If an AI helps me to achieve visibility, then I would already say that it might be more than this mechanistic claim “I’m not doing a blog post”.
Mirko Lange: So absolutely right. We’re not controversial at all. What I’m just trying to tell people is: go beyond visibility! So that was what we call input, output, outcome and outgrowth in impact research: Visibility is first of all only the output. This is, of course, the prerequisite for anything to have any effect at all. But my goal or my motto is always “from reach to depth of reach. Because we already have a trend towards increasing reach and visibility, but insanely reducing the depth of reach, i.e. the famous eight seconds or one millimeter depth. And that’s where we have the danger of commodity, that people surf along the surface so much that there’s no penetration at all, penetration into the head, that somehow a change, a change in attitude takes place.
And that is ultimately the question of strategy. So a Fast Moving Consumer Goods manufacturer naturally has a completely different objective than any B2B manufacturer who lives from competence. Because if my profile is based on competence, then range and superficiality are of course not particularly helpful.
Part 3: From digital content to synthetic content – what is the future of content?
Roland Fiege: We’re talking about a subject area that was still in its infancy twelve months ago, 18 months ago. And what has happened in this direction in the last year alone is unbelievable.
Let me ask you two separately, what do you think we’ll see when we sit down together in a year or two years’ time? In the context of content, AI, and yes, scalability, will we have any chance at all to go more in depth again? Or will all the topics just be surfed on the surface and dealt with at the headline level?
Peter Gentsch: I’d be happy to start. Once we have now spoken very strongly of text, which we see in AI development. But you also have to say that, of course, images and moving images are also increasingly being generated. And that’s incredibly impressive. Two or three years ago, no one could have imagined that I would simply enter a prompt “Teddy bear paints a portrait” and suddenly I get a video. That’s insane. I believe that the whole thing will be multimodal, that content won’t just consist of text, but that I can use different things.
The models, the so-called foundation models, are developing exponentially! Elon Musk’s GPT-2 was a huge model just a short time ago. In the meantime, it’s kindergarten. That means we’re going to see a very brutal development. And these algorithms practically access digital knowledge that is available on the Internet. As we all know, the Internet of Everything is becoming more and more digitized, which means that the amount of data for the algorithms is increasing, and the performance of these algorithms is rising.
That’s why I think the quality is getting better and better and also the possibilities to create different formats. But that doesn’t change the basic mechanic that at the end of the day these systems crunch zeros and ones. I say the illusion of creativity is getting bigger and bigger. The illusion, it’s like it’s created by a human, it’s getting bigger and bigger. The basic mechanics don’t change.
And I think what’s going to be even more exciting is, of course, the evaluation issue. Because these algorithms do nothing other than recycle, reproduce things on the Internet.
A very simple example: If you type “women can’t” into Google today, the auto-complete usually suggests “can’t drive” or “can’t park”. This means that the algorithms represent an opinion that is held by a large majority on the Internet. So the question is, who actually evaluates this? There are dozens of studies that say women can park better than men! But apparently the buzz on the Internet tends to disagree. And these algorithms are learning, I’ll say, mainstream. And so I see this evaluation component as insanely difficult, because we’re increasingly generating content that reflects a general opinion and just doesn’t give you the opportunity to intrude, as Mirko says, to position yourself. So we’re getting a little bit soft-pedaled there, I think.
I would say that now it’s not just about writing a blog post with AI. I think we’re already at the stage where we’re moving from digital content increasingly to synthetic content. And I don’t know if I want to live in a world where we only have synthetic content. But we’re well on our way to that because it’s so insanely efficient. An AI works 365/24, has no works council, never gets sick. And that’s why my prognosis is not so positive, because I think the idea of efficiency will drive us into this synthetic content bubble. I don’t think that’s necessarily a good thing, especially in terms of the quality of content. But again, there are new algorithms coming in at a weekly innovation rate. That’s gigantic.
Mirko Lange: I think what you just said is sensational. So the term synthetic content, I think that’s very, very good.
Roland Fiege: You should actually steal this one.
Mirko Lange: He has already been appropriated: Copyright Peter Gentsch. Also with the imitation. At the moment, we are also seeing that we have more and more competitors. They’re no longer copying us; they’re pretending to – by first imitating what’s visible from the outside.
I mean: children learn by imitating, that’s fine, that’s their first step. But this deep understanding behind it, what’s not there: I think we can make five more podcasts about that. Also what the social effects and backgrounds are. That people are perhaps also overwhelmed by the complexity and therefore don’t want to delve deeper into the matter.
What you just said, Peter: That’s exactly the point. I hardly dare to say it because this has been discussed so widely and very controversially. I’ll mention Precht/Welzer now, whose book certainly has weaknesses. But that’s the key point: they said they agree on a narrative. But what you just described, that’s exactly it: That the media agree on a narrative, because they also report more and more algorithm-driven and report about what people want to hear.
They’ve called that unwise, because nobody wants to hear “Gleichschaltung” because the term has negative connotations with fascist states. But many people don’t want to believe that this necessarily has to happen because of digitalization, synthesization and algorithms. And Peter, I’m completely with you on this, and perhaps also a bit old fashioned: strategically, we have to work much more, we have to use data. The way people use data today is actually a blind spot. My thesis is that we only understand a My of it at all.
So data-driven communications is a buzzword that’s electrifying the industry right now. But if I have data and I don’t understand what the data is really saying – or have incomplete data, then that’s the biggest bullshit I can possibly do. Or it leads to big nonsense.
Podcast anhören:
Teilnehmer
Mirko Lange, CEO, Scompler
Mirko Lange, Scompler
“Content Marketing Papst”, “der Deutsche Joe Pulizzi”, “Content Strategie Guru”… wenn Mirko Lange auf Konferenzen als Redner vorgestellt wird, dann finden die Laudatoren viele nette Begriffe.
Mirko Lange ist einer der renommiertesten Strategieberater für Content und Content-Marketing im deutschsprachigen Raum. Er ist Gründer und Geschäftsführer von Scompler, einem strategischen Redaktionsplanungstool mit mehr als 20.000 registrierten Usern und rund 150 Unternehmenskunden.
Prof. Dr. Peter Gentsch
Prof. Dr. Peter Gentsch
Prof. Dr. Peter Gentsch ist Speaker, Unternehmer und Wissenschaftler in einer Person und zählt seit den 90er Jahren zu den Pionieren und Top Experten im Bereich Digitale Transformation, Künstliche Intelligenz (KI) und Big Data. Während andere nur über Digitalisierung und Technologie-Trends referieren, lebt er sie.
Mit zahlreichen Unternehmensgründungen und -beteiligungen und fünf erfolgreichen Exits gehört er zu den erfolgreichsten Internet-Unternehmern in Deutschland.
Er redet nicht nur über die digitale Transformation und KI, sondern prägt und gestaltet sie. Dieses Wissen und seine Erfahrung vermittelt er als Coach und Speaker in Unternehmen.
Roland Fiege (Host, Managing Director rellify Deutschland GmbH)
Roland Fiege
leitete viele internationale Unternehmen als Senior Direktor im Bereich Social Strategy und Geschäftsführer. Nach seinem MBA an der Rhein-Necker Universität in Mannheim hat er im Bereich Social Media Balanced Scoreboard (SMBSC), Social CRM und Analyse Systeme geforscht. Er unterrichtete in den Bereichen Strategie-Entwicklung für Social Media, Monitoring & Analysen an verschiedenen Universitäten. Bei rellify bringt Roland seine Erfahrungen und Leidenschaft für Marketing, Unternehmertum und Technologie als General Manager EMEA ein und leitet das rellify-Team in Europa.
Roland Fiege ist gefragter Gesprächspartner verschiedener Nachrichtensender und Rundfunkredaktionen, wie z.B. RTL/ntv, ARD, Deutschlandradio und anderen. Gemeinsam mit Karl-Heinz Land publiziert er den bekannten Podcast Erde 5.0.
Peter Gentsch: The difficulty of the discussion is nicely illustrated by what Google said the other day: They’re banning AI-based content. They will filter it out. Roland, I think you know a site that investigates whether content was created by humans or not. The only problem is: I don’t recognize it. And the louder Google trumpets that, the clearer it becomes that they just can’t do it. And I think that also shows a bit of helplessness.
Mirko Lange: Even if the imitation is so perfect, it remains an imitation. That has to be studied sociologically. The variance is also going down. I see that as a big problem. If we look at diversity, we have so much parallelism now, for example in the gender discussion, but some people say there are only two genders. Evolutionary researchers say “evolution and life consists of diversity!”. If the AI learns at some point to produce diversity and to consciously give away errors, i.e. also to imitate defectiveness, then I am curious to see where this will lead.
Peter Gentsch: I recently gave a lecture to creative people and there was also the question “Imitation, what is that actually?” One of them said this beautiful quote: “Good artists copy, great artists steal.” Along the lines of “Well, that’s common practice. And every journalist does nothing else than recycle content on the Internet”. Some of them were very relaxed. But again: the danger is when, as a consumer, I don’t recognize something as an imitation – then of course we have a problem.
Roland Fiege: I can recommend “Data and Society” on this topic, i.e. the website datasociety.net. They deal with studies on the sociological or social implications of data-centric technologies and automated communication. There are already very, very interesting initial studies on the subject. Among other things, I also found the website haveibeentrained.com in an article there, where you can supposedly see whether your written content or pictures have already been used to train and perfect AI algorithms.
And then of course we come to the topics of “copyright” or “right to one’s own work”. Where that will be in the future when everything is remixed has always been an internet topic anyway. But who do we sue if an algorithm reuses content and builds something new out of it? In the past in music there were big scandals about sampling and also lengthy processes. I think that’s the end of it. Ultimately, it will no longer be possible to differentiate between synthetic content – Peter, I love the term – and originally written content. But as far as the purpose – and this is actually very, very similar to people’s leisure time behavior at the moment – is concerned, the actual sense and purpose and the enrichment of meaning, I believe that the person should primarily be able to manage and also decide. Because only then can we make a difference. Otherwise we have really softened and smoothed-out content that consists only of jumping from one headline to the next.
A little anecdote: We are also a purely technology-driven company with lots of data scientists and programmers. Everyone works remotely on the computer all day. And then in the last social call we asked “Yes, what are you doing? What was the best moment in the last few weeks or on vacation? Or what are you doing on the next holiday? Weekend?” And nobody was there who said, “I’ll put on VR glasses and go to Decentraland and stay online.” Instead, everyone said, “I’m going hiking,” “I’m going to do some handicrafts with my children ‘ or anything. But nothing on the computer. And I think there is a bit of truth in that at the end of the day, that of course we use the technology, and we do everything that is possible. Anything that can be built will be built and will be used.
But in the end we also have the purpose and we feel it in us. And that’s challenge enough to always link that to organizations and commercial organizations. I think that’s where our strength lies and the AI won’t be able to take it away from us anytime soon.
Peter Gentsch: Although, in the discussion with the creative people, I found it very exciting that they said, “Well, somewhere there is also a democratization of creativity”. Mirko, we got to know each other in the area of social media and there we also have a democratization of the media. Anyone can send and receive. And now I can actually say: “Everyone can take great pictures”. I’m a complete mess artistically, but all of a sudden I can prompt something and get a great picture or a great video.
You could say: Then we will all become creative people and innovators. Some people probably don’t want to hear that now. But I also found the idea quite good when we talk about AI in such a negative way. You could also say, “Hey guys, everyone is becoming an artist now. A democratization of creativity.” I know, Mirko, you will probably vehemently disagree, but I found the idea very interesting.
Mirko Lange: No, no, I don’t want to contradict at all. But as I said, I’m now interested in the sociological consequences, because that’s how creativity becomes a commodity. But creativity is therefore only a good, like sex. If it’s a commodity and it’s readily available, it gets boring. Well, I always make the stupid joke when people ask me because I’m a passionate golfer: “Are you still having sex or are you already playing golf?” And then I also say “Well, I can have sex anytime, but a great one Don’t drive.” The availability of goods and unlimited availability, which is the definition of a commodity, lowers the value.
Roland Fiege: Yes, totally.
Mirko Lange: Only through scarcity do we have values. These are very fundamental rules of life. So is it even desirable to democratize creativity? What’s in it for us? Nothing! The meaning of creativity will eventually disappear!
Peter Gentsch: I totally agree with you there. At least I had small moments of happiness when I produced some videos here.
Mirko Lange: Absolutely! In the short term yes! That’s it: we have our reward system firing. Dopamine is released and so on. We’re like drug addicts, even in social media – it’s always just about the little dopamine kick that we’re all addicted to. But with every addiction, the dose has to get bigger and bigger and the limits of growth and so on… I find that an exciting topic. Where is this supposed to lead?
Actually, I still think “HI” is the most beautiful of all. And I find this trend, i.e. “human intelligence”, this trend the worst, that people are finding it more and more difficult to think.
Roland Fiege: That is correct. But I think we’ll do the next podcast about that.
Mirco Lange: But I think that’s what makes being human.
Roland Fiege: Yes, of course!
Mirco Lange: What’s in it for me? When people only eat fast food and let the AI think for them? Then I somehow optimized the dopamine release, but it doesn’t lead to happiness, definitely not!
Roland Fiege: Then I have a suggestion for goodness at the end. So synthetic content that is consumed and perceived by synthetic consumers, who then synthetically consume and …
Mirco Lange: And get a synthetic feeling of happiness.
Roland Fiege: No… and then we go golfing… And then ultimately we have our purpose there and say okay, the machines will do the rest. And then we realize: “It’s actually super interesting what’s happening right now”.
Peter, thank you for your assessment. I’m super excited to see what quantum leaps will be made in the coming months. We will observe that, of course we will also communicate and share it here again and again in the context of this podcast.
And many, many thanks Mirco for your expertise, your assessments, including the philosophical ones, which are important. And I think we’ll meet again, we’ll see each other again. Thank you both so much.
