AI – friend or foe?

For many people, concepts such as artificial intelligence and machine learning are inextricably linked with uncertainty and fear. Does artificial intelligence mean that robots will assume all power? And will everything be automated so that in the future only programmers will have jobs?

Bure-Noréus.jpg

Bure Noréus, Data Scientist and AI expert.

No, real life isn’t actually that frightening. The development of artificial intelligence – commonly known as “AI” – can be compared with the production of steel. They are both general technologies that can be used for good or bad purposes. But steel – or AI – is nothing to be afraid of in and of itself.  It is possible that some professions may disappear on account of AI, but this is not a unique development. Over the past 200 years, professions in a variety of different areas have been created and wiped out on account of technological innovation. 

“In ages past, almost everyone worked in agriculture. Due to technological innovation and automation, only a small fraction of the labor force works in agriculture today, but that doesn’t mean the rest of us are unemployed. In the same way, you can expect that AI will change the labor market once more,” says Bure Noréus who, in his role of Data Scientist, has been working with AI and machine learning for years. 

Both AI and machine learning are set to make our lives easier in a number of areas in the future. One example is driverless cars, which are likely to mean improved fuel economy and fewer accidents on the roads.

“It might sound like pure science fiction, and it may still be a little way off, but it’s not beyond the bounds of possibility that our children might not need to take their driving tests,” says Bure Noréus.

Already exists in society

In the short term, AI is likely to generate incremental innovation in the form of minor improvements in efficiency in a range of areas. AI can recommend songs and artists, making it easier to find music you like. AI can fine-tune internet search engines. AI can predict when your mail parcel won’t arrive on time. In many ways, AI already exists in society, improving our lives in ways we may not necessarily notice.

In the slightly longer term, AI-assisted automation may produce major benefits and result in many things becoming cheaper. The cost of labor is often the most expensive factor in a production process, and if you can automate the processes that are most labor intensive, you can reduce costs and therefore sell your products more cheaply. 

“AI and machine learning produce the biggest benefits where access to data is easiest,” explains Bure Noréus. 

He identifies internet commerce as one of the most obvious areas. 

“Huge volumes of customer data have been collected in webshops. These data generate a deep pool of knowledge about consumer behavior. What do we buy? When do we buy? And why do we buy?” 

Diagnosing illnesses

In the health and care sector, AI can also be a useful tool in the context of diagnosing illnesses at an early stage. Here, however, it will take much longer for AI to play a decisive role; given that an individual patient is required for each data point, it takes a long time to collect data. 

As AI learns from data collected from real life, a “cheat sheet” is needed, too. If you want to train an AI system to diagnose brain cancer, it’s not enough simply to feed the system with images of people’s brains – you also have to tell it which brains had cancer, and which did not. This is how the AI system learns. 

In this case, the “cheat sheet” is expensive to create because it requires a brain specialist to tell the AI system what is right and what is wrong. In e-commerce, on the other hand, little if any manual input is required to create a “cheat sheet.”  The information about who has made a purchase, where and when has often already been saved.