“Hey, car, take the children to their piano lesson – I have an important appointment with a customer!”

In our first interview with Oliver Kramer from the University of Oldenburg, we talked about the basics of artificial intelligence (AI) and provided some clarity on the topic’s biggest question: what exactly is AI?

Now, we dive further into the topic with more questions. How will AI assist us in everyday life? What are AI’s potential applications and will machine learning change our professions? Oliver Kramer not only answers these questions, but also explores their moral and legal implications.

Hello, Oliver! Thank you for taking the time for a second interview. Today, we would like to go into more detail and find out how AI could change our daily lives. In your opinion – as an expert in the field – how will my job (as a content manager) change over the next few years as a result of AI?

I think that we have to think more about data, and what can be learnt from data. Wherever processes can repeat regular activities by learning algorithms, the greatest changes will be felt.

It depends on the activity. Some working environments will certainly change quite significantly. For example, AI will increasingly aid working processes using augmented reality. This means that the AI can identify what’s going to happen next. For example, consider the process of building a piece of furniture from Ikea but, instead of reading the instructions, you put on a cool pair of glasses that show how to fix it together, or which tool to use next. This means that AI enables us to perform tasks for which we might not have been trained, such as assembling complex components.

And that is actually a big advantage. People often criticise AI for destroying jobs, but it can also create jobs – not only for AI programmers, but also in fields such as the one I just described. Because now, someone who is unskilled can put on a pair of glasses, find out exactly what to do and learn from the process. So, this creates new opportunities for unskilled workers to perform more advanced tasks without having been trained to do so.

AI supports. And, in principle, this can be applied to all other occupational fields, including those of the professor, the researcher and the data scientist. We have to use as a tool this ability to perceive things beyond our understanding and learn from data – we can do that now, thanks to AI and learning systems. We can read huge amounts of data in one go, make use of new sensory abilities and think in other dimensions. We have to make sensible use of these features and integrate them into daily work routines.

What this might mean for an individual’s work process varies. For example, AI could improve a text I’ve written because it can learn effective writing techniques from millions of existing texts. So a feature of AI can improve texts, including journalistic pieces.

… Maybe AI can even write its own texts?

Writing independently – exactly! I provide the AI tool with just a few key words and that’s enough! Or, if you’re conducting an interview (like we are right now), AI could produce an automatic transcription. In terms of recognising the spoken word, appointments could also be entered directly into a calendar, notes could appear in bullet-point format on my screen, or automatic searches for suitable documents could take place. You can imagine so many things that are possible. The more difficult decision is choosing what is most urgent, right now.

“You can imagine so many things that are possible with AI. The more difficult decision is choosing what is most urgent, right now.”

This means that, at the moment, you can’t say what will be possible next?

Yes, there is so much we can imagine, but choosing the smartest idea and turning it into a product is the interesting thing.

So, if I understand correctly, you don’t have the dramatic expectation that we will soon all lose our jobs to AI?

Absolutely not! We should instead focus on learning to use AI as a tool that complements our work. I think that completely replacing work doesn’t make much sense, because AI cannot replace the human. It will not replace human beings and is not able to replicate human decisions in its processes.

Many of us currently make use of language assistants, such as Siri or Alexa, more or less successfully. Do you think this area will evolve further? Will we all soon have a personal, smart assistant?

Yes, that’s one of the next big developments. This is about so-called ‘Horizontal AI’. This type of AI supports as broad a range of tasks as possible: accompanying us in everyday activities, recognising our language and supporting us in making decisions. For example, this AI might remind me that I shouldn’t drink any more coffee today, because my caffeine consumption is already too high. This type of AI can, if I wish it to, influence my life and help me make smarter decisions. So this is definitely a field where a lot is going to happen and a lot of money will be made.

However, the big players are already on board with this, so Google, Apple and co. are the first to sell us such products, because we’re already constantly running their software.

Aside from Google and others, can you name some other companies already making use of AI?

At university, for example, we use AI to predict data pertaining to sources of renewable energy – such as predicting how much wind will blow through a wind farm. This can be executed using weather models, which are also data-based. We use such processes in research, but there are also companies selling them. The intended consequences of all this are that we waste less energy and are able to respond to a very dynamic energy environment. In the past, there was one power plant that supplied energy to everyone. Today, there are decentralised suppliers everywhere, but there are also decentralised energy consumers. In order to somehow create balance in this field of potential tension, you can use AI.

This type of AI can show when an area of high wind moves over Germany, when it is cloudy or very sunny – important for solar energy. But it also indicates the best times for me to start my washing machine, or when electricity is at its cheapest.

Beyond that, autonomous driving is of course a huge field, from assistance systems to fully autonomous driving. There are lots of start-ups working in this area.

The topic of autonomous driving isn’t too far from impacting our everyday life, is it?

No, not far at all. In Palo Alto, I’ve already been overtaken by autonomous cars created by Google. And all other major car manufacturers already have concepts ready.

So will autonomous driving be one of the first AI topics to really reach us?

Yes, definitely. This topic has its legal problems, however. Who bears responsibility if an accident happens, possibly leading to personal injury? Is it the responsibility of the AI programmers, the data supplier used to train the AI, the manufacturer, or the driver? AI scientists think a lot about this issue – the moral aspects.

The MIT Moral Machine, for example, forces you to decide who to sacrifice in the event of having to make a critical decision – a really unpleasant thought (you can test it yourself online). So, in a dangerous situation, who is most at risk? And this brings us to the point that such moral decisions should not be left to AI alone. In this area, there must be a shared conversation, as well as a legal discussion, about how to make the ‘right’ decisions. This will definitely be an important topic in the near future.

But you should also consider the potential benefits. Autonomous driving will be more energy-efficient. You’ll be able to share cars that travel from A to B, taking one person to their destination and then the other. It will also be possible for cars to do things they currently can’t: “Take the children to their piano lesson – I have an important appointment with a customer,” or, “Fetch Grandma from Munich!” These abilities will make a car completely independent, which will give us a completely different understanding of mobility.

Photo by Andy Kelly on Unsplash