Can we trust AI and what are the benefits for businesses?

With all the ongoing discussions about artificial intelligence (AI), we decided it was time to get the viewpoint of our own AI expert, David Degerfeldt. David who joined the company this year, has worked on and launched multiple AI projects in diverse industries. So, what does he have to say?

David Degerfeldt, AI Expert at Data Ductus
Photo: Jonas Westling

Tell us about your background in AI?
I first started working with AI-related projects at Ericsson, where we were developing tools to predict speech quality. However, three years ago, I took on the responsibility of rolling out AI at the Boliden Mining company. I was responsible for introducing an AI program into the organization that would improve production efficiency and drive digitalization. This involved identifying the best use cases for AI, deploying it, and using the results as proof of concept. The mining industry is conservative, so we had to show the organization that this technology shift was the best way forward. It required both cultural and technical change.

Can you give us an example of an AI project that made a big impact?
Absolutely, and it’s important to mention that we weren’t looking at administrative tools such as RPA but at production, where small improvements can make a big impact on the bottom line.

For instance, Boliden has a large-scale flotation process that separates valuable minerals from waste material. This complex, unstable, and difficult-to-control process involves numerous interconnected tanks, each the size of five shipping containers. Previously, it was controlled, to a large degree, by human operators who tweaked flotation conditions based on reactions in the tanks.

We developed an AI solution that could always consider all available, i.e., even historical, data. The AI tool has learned the Boliden flotation process, analyzing the different actions performed by operators based on reactions, and now feedbacks to the operators advising when and what are the best actions to take. This has led to improvements in output quality and delivered production efficiency.

And now you’re at Data Ductus?
That’s right. I decided to make a change before becoming too attached to the projects I had implemented. And besides, I wanted to try working as a consultant where I could share my knowledge with many other businesses to help them improve their operations. There is a lot of uncertainty at the moment around AI, but when used properly, it can offer significant benefits – and I’m looking forward to helping organizations implement them. For instance, Ductus is already contributing to the green industrial revolution taking place in northern Sweden, but with the right AI solutions our efforts could be increased.

So, you are very optimistic about AI?
Definitely, I’ve spent a lot of my career working with it to improve processes. Also, I think one has to be because AI will play a central role in our futures. It’s up to us to make sure it’s the proper role. For instance, it’s already detecting cancer in patients at an early stage – AI systems have been taught to recognize the signs of cancer in images that we, or rather a doctor, wouldn’t necessarily spot. And it’s not replacing doctors but rather acting as a support tool for them. AI is also being used to develop molecules that will help treat future sicknesses. I believe it may even help us solve the climate crisis since we don’t seem to be able to do that ourselves. And for the companies we work with, we can help them be more efficient and effective.

You have to understand AI for what it is. It doesn’t have feelings or wishes; it’s just a fast learner based on scanning large data sets.

What about the fear factor of AI?
We generally fear what we don’t know or understand. ChatGPT, to name the most talked about AI tool, has dropped so quickly that it’s caught the world by surprise. Some countries have gone as far as banning its use, although how they’ll do that I’m not quite sure. Like many AI systems, it’s a tool that’s been trained using large data sets with fairly standard algorithms. It’s the data set that is impressive, much of the web, pre-June 2021. With that being said, it can’t detect what is fake news; it’s up to us to do that. And, then there are deep fakes, that many of us have seen examples of. As a consumer of information, AI can make it difficult to distinguish the truth.

Personally, I believe we will see new regulations coming into place shortly, similar to the new environmental regulations that are being rolled out in Europe for companies to report their emissions and impacts. The fact is, AI is moving faster than we can. It can offer great benefits to society if put to the right uses. It’s up to us to embrace it and benefit from it.

About David Degerfeldt.
David has been working with AI for well over a decade in the telecom and mining industries and has a Ph.D. in Computational Electromagnetics. When not working with AI he likes to spend time in the great outdoors, and is a self-proclaimed home automation junkie, developing complex solutions to solve simple problems.

Email: david.degerfeldt(at)