How AI will have an impact in 2024?

Mikael Munck: Hello and welcome to AI Watch. My name is Mikael Munck and I’m the founder and CEO of 2021.AI.

Today I am honored by having Henrik Fabrin from Dansk Industry with me. Maybe, Henrik, you could give a short introduction of yourself.

Henrik Fabrin: Sure, I’m Henrik, I’m the AI lead at Danish Industry. I’ve been there for three or four months. Before that I spent twelve years as an entrepreneur like you. Founder, CEO, building, taking AI products, companies, had a lot of fun there. And then I wanted change. Then I thought, as per our conversation, AI is going to impact everything. So where can I have a huge impact on helping Danish companies? And I can do that through the Danish industry. So that’s pretty awesome.

Mikael Munck: That’s fantastic. And that’s actually one of the reasons why we invited you to come here today, because you both have the background from the entrepreneurial world, successfully running startups, and then now engaging more with, let’s call it, big classical industry. What do you think will be the biggest impact? Where is it that we’ll see AI have the most influence in 2024?

Henrik Fabrin: That is certainly a very big question.

Mikael Munck: Yeah, it’s a big question.

Henrik Fabrin: I think there’s many viewpoints on it. I think one of them is, if we zoom in on, just for the sake of it, generative AI. Right, so 2023 was ChatGPT and everything. Large language models became very easy to use, so the adoption went through the roof. That cost nearly every company to adopt it in some form. Either bottom up by the employees or the leaders, had a strategy, but also put a lot of different conversations and people thinking about everything from how do we use it safely, how do we not use it, where is it going? Is this just going to ruin trust on the Internet? How can we use it to win, to enable new business models, how can we keep what we’re doing?

Companies and people being more aware and thinking more strategically of how is this going to impact our industry? How is this going to enable us to maybe do something new in Europe? Shaped within the box of the AI act, which is going to be in place.

Mikael Munck: Now, one point that a lot of organizations have been discussing in Europe now is with the EU AI act and with the responsible use of AI and that whole debate. Do you think that actually will be a hindrance for the members of Dansk Industri or what is your view on this?

Henrik Fabrin: The members of Danish Industry are pretty broad. So it represents Danish society. So 50% of members have less than 50 employees. And then of course, you have a lot of big corporations.

I think especially for the smaller companies, SMEs, it has a lot to do with what do we even do? How can we take this opportunity? We have limited resources, limited time. How do we do it in an easy, fast way where some of the larger companies, they have a large organization, they have maybe a set up where maybe they can take the time a little bit more, but they can also be thinking a little bit longer because the organization allows them to do that.

Mikael Munck: So it is something about investment and the resources that goes into this from a lot of the organizations, for sure.

What about talent? What about having enough people who really understand more of the core of this? Because as you rightly said, just about everyone now has been in touch with AI. Of course, on their mobile phone, but maybe also via ChatGPT and others. But then the challenge is how do you transform that in, or how do you integrate that into your existing organization, your existing IT infrastructure and so on and so forth?

Henrik Fabrin: Again, large corporations can have an advantage because maybe they have more resources to adopt, where smaller companies have to rely more on standard platform, standard technology. And then we did some research together with some other organizations, and we saw that in the Nordics, 50% of organizations that were already using AI, only 15% had an actual strategy. And that, in my view, says that it’s become a lot easier to use, to use knowingly in your organization. So it becomes much more a question for the leadership, the CEO, the CIO, the CFO, how do we enable that and how do we actually transform our organization?

To think about AI first, I’m a big believer in an AI assisted organization. So everyone from CEO to intern will use different types of AI, different types of models, depending on the different tasks they have throughout the day. And they will also bake AI into their products that they service to their companies.

Mikael Munck: While you are now a startup founder, I know you work with AI. I know you had to make your own AI strategy. Think about how you got it into your product. How did you do that successfully?

Henrik Fabrin: Two directions there. One is, how do you integrate into your organization, right? And then how do you bake it into your product? My company used a variation of some of our own core stuff, and then we used open source or standard software. Don’t marry too much with one model provider, one service, one vendor, one type of technology, because the innovation, and it’s going so fast, as you know. So who knows where we are? What will be the dominant technology in 6, 12, 18, 20 months? Do it in a way where you can easily plug in and plug out other providers. So you keep your options, but you get the learnings.

Mikael Munck: And I guess that’s of course from a startup/scale up founder point of view. But I would actually say the same would go for the large organizations also, because I believe that it’s very hard to predict, maybe even if there will be a dominant player in this, what the next AI model will be doing.

Just one week ago now we saw a new model from OpenAI doing videos in a format that we did not believe were possible and in a length that we did not believe were possible. If we just go back some months and I guess that’s the dilemma you’re really saying you have to be open, you have to be flexible in the way you work with this.

Henrik Fabrin: Exactly. We need to have different types of connectors so you can easily switch out, right? Both in terms of cost model, size of models, where is it coming from? But also just stability wise because we’ve also seen that some of these are, the adoption is so fast that it’s difficult for them to really keep up. Even if you’re like here in Nordics, Microsoft is very dominant, right? So many large organizations have Microsoft all the way through the product, which is great, but still there. Use it in a way where you have options so you can, especially on models. Use closed source models, use open source models, use different types of models.

Mikael Munck: With this, let’s say flexibility. With this openness, with thousands of models potentially in production, I guess there must be some challenges to ensure that you’re compliant to regulations and ethical guidelines. And what else we’ll see that we need to comply with as organizations.

Henrik Fabrin: 100%, right? AI will lift everyone, the good actors and the bad actors. If we disregard the bad actors and then talk about good actors, everyone wants to do it in a good, proper, safe, secure, responsible way. And the better the tools are to help them do that. Easy, then you could do that. If it costs a zillion dollars to do, then it’s only the last corporation, then the SMEs are screwed, right? There are a lot of challenges in that. You can slice it into some layers. There’s the guidance, the advice, how to do the right approach that consultancy firms and lawyers will help do and translate the law into what to do. And then there’s the product, the tech layer. So using tools like 2021.AI and others to say, okay, how can we do that? Deploy it, measure it, and report on it in a way that’s easy and doesn’t take 20 data scientists and an IT department of 200 people.

Mikael Munck: And I guess that’s also the innovation we need to see. I see very often that there’s a big gap between being a tech-oriented company that understands everything around infrastructure, flexibility and all of that and then being a less technical-oriented company that might only be on a journey towards digitalization to some level within the organization. And all of a sudden you have AI on top of that.

Henrik Fabrin: And I think one of the key things here is that most of them are already using it. They just haven’t figured out exactly where and how to use it long term. And this comes on top of everything else that’s happening in the world. So no wonder that the senior level in the companies are a little like, “what are we going to do”, right? Because they have everything else to focus on as well.

Mikael Munck: With that great remark to end this conversation. Thank you very much Henrik for this. Thank you.

Henrik Fabrin: Thank you for having me.

Watch the previous episodes