Inside Robovision: “Within five years artificial intelligence will have become omnipresent”
Jonathan Berte, who trained as a civil engineer, smiles as he thinks back to the pioneering years at Robovision. “In fact, when I was a kid I had a really analytical mindset. In the scouts and at school I used to keep note of absolutely everything. It was really important for me to collect information. I was a kind of ‘infoholic’. But just gathering information gets you nowhere. That also goes for information that’s just stored on hard disks. The added value comes from using that information efficiently.”
How exactly do you do that at Robovision?
“Technology is evolving at lightning speed. These days just about everybody has a smartphone in their trouser pocket. Apart from anything else, these devices create a great deal of information, so we need to keep up on the algorithmic front and artificial intelligence helps us with that. That’s how we can provide governments, institutions and companies large and small with a platform for automated decision-making on the basis of visual data. In addition we constantly ask ourselves how we can democratise artificial intelligence. So in a way we’re like the Airbnb of artificial intelligence.”
What might that visual data be for example?
“In May, in collaboration with the University of Antwerp and security firm Securitas, we set up a smart camera in a shopping street in order to measure to what extent people were complying with social distancing requirements. This is important information for the decision makers in this country. Of course we don’t have to look through the images ourselves. We get them analysed using a specific type of artificial intelligence – self-teaching algorithms or what are known as neural networks. They’re designed somewhat along the lines of our own brains, though not nearly as complex.”
Which brings us to the fashionable expression ‘deep learning’. Are machines eventually going to make themselves smarter than us humans?
“Oh, that’s already underway at this very moment – in radiology, among other fields, plus also in games. Remember the legendary Go match between South Korean grandmaster Lee Sedol and a computer, which was beautifully represented in the 2017 documentary film AlphaGo? We’re also focusing on deep learning, because neural networks are very efficient at dealing with visual data. However, it will be some time yet before AI can equal a human being in intuition for instance.”
You’ve now evolved from a startup to a scaleup. Where do you want to be in five years’ time?
“The society of tomorrow will be one in which everything will be properly measured and dealt with. For instance, we’re also working in the field of horticulture, where AI can be applied in quality control – to spot fruit with an abnormal shape or colour, say. Lots of agricultural and horticultural businesses have got into difficulties over the last few months because pickers from Eastern Europe weren’t able to enter this country. Those businesses will very probably be investing in AI and automation over the next few years. In these kinds of fields, the coronavirus has taken us to a digital society almost overnight.”