The world of digital marketing is moving fast – and that includes AI. That’s why we’ve invited Rune Jensen, Group CTO at Generaxion and AI expert, to the studio to delve deeper into the topic. And to answer the big question:
What can AI do for digital marketing in 2025?
Take a deep breath and delve into the topic with us, in our mini-podcast Generaxion AI.
1. What defines artificial intelligence?
Within AI’s complex machinery, we find three equally complex layers:
Machine learning : the process that gives computers the ability to mimic how humans use information to learn and improve.
Shallow learning : machine learning with few processing layers, used for models intended to handle quick and simple tasks.
Deep learning : machine learning with many processing layers, used for models that handle more complex and data-heavy tasks.
But when we lift the hood and expose the AI engine, another crucial question arises: is what we call AI really AI?
With his twelve years of experience, Rune gives us insight into the AI winter, machine learning, how AI should be defined – and why ChatGPT was a certain breakthrough.
Episode 1
2. Large language models
The real breakthrough for AI came with the emergence of large language models (LLMs) such as ChatGPT, Google Gemini, and (most recently) DeepSeek. These models are all based on natural language processing – one of the most complex disciplines in machine learning.
This is what makes it possible to give advanced prompts and even have conversations, as the models seem to constantly learn from every input you give them.
But it turns out that despite their complexity, the models can’t actually get smarter.
Many people think that just by using ChatGPT it will become more intelligent every time. But that is simply not the case. They are static models. They do not become more intelligent – they get a little more knowledge, more prompts, but they do not become smarter.
While this may take away some of the magic of AI, there is a very good reason why this is the case – one that Rune can tell you more about.
Episode 2
3. Development and limitations of AI
As AI evolves, the scope of what it can accomplish seems to be growing exponentially. But while there is undoubtedly enormous potential in AI tools as a whole, there is much that AI cannot yet do. Rune highlights two areas in particular:
First, AI has no self-directed purpose. While it has an intentional use case defined by its developers, it has no inherent desire to learn and improve. And in some areas, that is precisely what is necessary.
Secondly, as users, we demand accuracy – and while AI can deliver this to some extent, ensuring accuracy in more complex calculations requires large resources. Resources that we are still implementing in models like ChatGPT.
Episode 3
4. Human vs. Artificial Intelligence
“The robots will take our jobs.”
An old and persistent fear.
It’s a discussion that’s still ongoing, even as we get used to AI tools becoming part of our everyday lives. And Rune’s message is clear:
You want AI to support the work you do. You don’t want it to take over.
The reason for this is all about what humans can do – and that AI simply can’t. From ethical considerations to questions of quality over quantity when working with a machine that is basically just programmed to create the result you asked for.
What would it take for us to even consider putting responsibility and tasks entirely in the hands of AI? Do we even have the technology to make this possible today? And if not – when will we have it?
Episode 4
5. Digital marketing and AI
So how can AI be used in digital marketing in 2025?
AI is already present in most marketing tools, but (not entirely unexpectedly, given what Rune mentioned in the previous two sections) it does not yet have the ability to replace human quality. However, there are areas where AI can already be incredibly valuable to implement:
- AI is significantly better than humans at math and calculating statistics, so using it to, for example, predict the outcome of a campaign can be very useful.
- Tools like Metas Advantage+ make it easier to set up target groups for your advertising campaigns – with a hit rate that is likely higher than if you did it manually.
- Its processing power far exceeds that of humans, making it possible to analyze large amounts of data in a fraction of the time it would take to do it yourself.
In short: you can – and should – use it as a support tool.
Want the slightly longer version? Listen below.
Episode 5
6. The future of AI
Can we expect robots to take over the world in the future?
Who knows. But it’s not a reality waiting around the corner.
It was only recently that we experienced a breakthrough with ChatGPT, and since then there has been nothing with similar capabilities. It is likely a long way to go.
You don’t have to worry about AI replacing your job – but your job will change because of AI.
And at Generaxion, we are always looking for opportunities to implement AI – not to replace anything, but to support it.
Episode 6