When hearing about artificial intelligence, do you think about robots taking over the world? Or, if you’re a bit more into the topic, about automated data mining and vicious suggestion- and advertising algorithms that undermine your free will? We are aware that in the context of media and information literacy, AI is often portrayed in a quite critical light.

But despite that being important, we want to take a step back from it here, and instead serve you our veg tech rice, which is carefully prepared with all the good and innovative energy that AI, one of the biggest achievements of the tech sector, has brought us.

But what exactly is artificial intelligence in the first place? The Cambridge Dictionary¹ defines it as the study of producing machines that can complete tasks like problem solving, learning, understanding language and recognizing pictures – in other words, AI is every instance where computers engage in behavior and thought processes that would otherwise require human intelligence. In the more specific case of deep learning, which is a subset of artificial intelligence application, algorithms can teach themselves how to best accomplish a set goal (for example to favor and suggest online search results that are relevant to you) by drawing from and continuously analyzing a related set of data (in this case, information on what you searched in the past and which search results you clicked on).

This offers an amazing range of opportunities. Among the ones that we encounter on a daily basis are the algorithms that learn our shopping or browsing preferences, or predict which music or videos we might like. Or the algorithm behind your smart navigation app, that helps you avoid traffic in real-time.

But there’s a lot more that can be done with AI: For example, the Canadian company BlueDot claims that its intelligent algorithm had predicted the outbreak of Covid-19 before the WHO². It achieved this by monitoring, among many others, data on how many international flights people booked, how medical and emergency personnel was requested and relocated across the globe and how the expression of emotions and the discussion of certain topics changed on social media. By automatically monitoring and combining a vast number of seemingly unrelated observations, the algorithm could accurately predict that something big and unfortunate was coming. In the future, similar prediction mechanisms could more and more help us predict, prepare for and ideally interrupt the spread of diseases.

Another project³ is currently aiming to create an AI that will be able to track global migration movements by analyzing censuses, administrative data, migration flows since 1991, as well as Google search- and air traffic data. Being able to predict where migrants are headed can be crucial information that enables humanitarian aid to be delivered to the right place at the right time. As one of the many AI advancements in the medical field, the company Envision, in collaboration with Google Glass, created AI-powered glasses that help visually impaired people “see” by reading text to them and telling them about the people they encounter, obstacles on the street, etc. out loud⁴.

Finally, we want to point out the algorithms that are widely used by the US police to predict in which areas crimes are likely to happen, like PredPol. In theory, this sounds very promising for making police forces’ work more efficient: they know where and when to concentrate their efforts on a certain part of town. However, the flipside is that this kind of software has been shown to perpetuate racial bias, because it relies on data of past arrests, which is, unfortunately, biased too: black people are five times as likely to be stopped without a cause and twice as likely to be arrested as white people in the US. Relying on this data, the algorithms mentioned often predict an unrealistically high likelihood for crimes to happen in black neighborhoods, resulting in more policemen patrolling in these areas, more black people being arrested and thus, the cycle continues.