Artificial Intelligence (AI) can make life a lot easier. “But if we do not put AI on a tight leash very soon, things will go horribly wrong,” critics warned last Friday during the conference AI for All – From the Dark Side to the Light at the Evoluon in the Dutch city of Eindhoven.
Experts, interested parties and various speakers spent the afternoon exploring how both science and society can counteract the dark side of artificial intelligence. “The dumber artificial intelligence is, the better it is,” contends even Bruce Sterling, sci-fi author, journalist and AI critic.
‘AI for all’, you would think
AI has many uses that make our lives just a little bit easier. We are already dealing with AI in so many places and instances in our daily lives. For example, uses for it are found in your cell phone or in your car. To a certain degree, we can no longer even live without it. For example, an AI system can also detect credit card fraud and the police use AI to track down criminals more quickly. In fact, even in the art world, AI is already wildly popular.
Although AI has plenty of uses that bring benefits, the speakers also saw inherent dangers to this form of automation. This is because AI systems rely heavily on massive amounts of data. The more data in the system, the smarter the technology becomes. Not surprisingly, the discussion here also revolves around data-hungry neural networks.
“Systems that open doors which solely rely on facial recognition technology are fast and secure, but this technology also has the ability to lock out entire populations, such as the Uyghurs in China,” Sterling points out. The elderly also increasingly struggle to adapt to new technologies. As such, the application of AI, in a sense, is creating a social divide. Because what is an advantage for one person may well be a disadvantage for someone else. Consequently, artificial intelligence is slowly but surely influencing our whole society; often without us even realizing it ourselves.
AI is not neutral
Take, for example, HR software that ignores female candidates in the recruitment of programmers. Or systems that start to exhibit racist tendencies. Nevertheless, this is not something that can be completely avoided. After all, the data AI systems work with is a reflection of reality. And biases are human. AI adopts those so-called biases.
“Which is why AI is increasingly coming to be seen in a negative light,” Sterling notes. “Some people call it a tragic technology for that reason. But others, on the other hand, see more opportunities than ever before.” For example, the speakers argue that companies that fully commit to this technology could see their revenues at least double between now and 2030. Companies that disregard AI will fall behind accordingly are projected to lose about twenty percent of their average cash flow.
Good design leads to misuse
During the panel discussion, the attendees wondered what will happen to their privacy if AI gains the upper hand. After all, artificial intelligence could very quickly lead to a society in which privacy has completely been eroded or even disappeared. Take, for example, facial recognition software in crowded shopping streets that can find out where you are hanging out in no time. Or that one app that you gave access to your camera to let it run in the background.
“By doing so, we are giving technology the ability to keep tabs on us, without us knowing what happens to our data,” said TU/e Professor Wijnand IJsselsteijn. According to IJsselsteijn, that kind of transparency is a very dangerous development: “AI is not explainable enough. Due to that lack of transparency, Big Tech companies profit from all the data that is collected while nobody has a clue what happens to that data.
“AI is very prone to abuse if it ends up in the wrong hands. That’s why they say it’s high time that AI is regulated. It’s time for lawmakers and governments to draw a line. AI can become as destructive as we ourselves choose. How far we let it go is mainly up to us.”
