Artificial Intelligence Needs Rules

It comes with the job that secretaries of commerce are rather dry people, as they deal a lot with dry topics such as growth figures, trade agreements, and subsidy guidelines. It is therefore all the more surprising that Germany’s Secretary of Commerce, Peter Altmaier (CDU), recently became downright euphoric during the presentation of his National Industrial Strategy 2030 when talking about artificial intelligence: "The applications of artificial intelligence," he said, "probably represent the greatest basic innovation since the invention of the steam engine." The politician is by no means alone in this assessment. But Elon Musk is not alone in his outlook either: The visionary entrepreneur and electric car pioneer (Tesla) considers artificial intelligence the greatest risk to our civilization.

The computer learns to learn

Olaf Groth is familiar with both points of view and can sympathize with either to a certain extent, but describes himself as "rather optimistic overall." An audience of more than 500 people had the opportunity to hear Mr. Groth’s expert take on AI at the Stuttgarter Gespräch, an event series co-hosted by the Robert Bosch Stiftung and Stuttgarter Zeitung, a daily newspaper. Originally from Viersen/Germany, he has lived in the United States for almost 20 years. One of the thought leaders on AI, he is an economics professor at Hult Business School in Berkeley/California as well as a member of the expert network of the World Economic Forum in Davos/Switzerland and head of Cambrian Group, a management consultancy that specializes in artificial intelligence.

For years, artificial intelligence has been praised as ‘the next big thing’ among technology experts. Nevertheless, it made perfect sense that moderator Andreas Geldner, business desk editor at Stuttgarter Zeitung and a start-up and innovation specialist, asked first for a definition of artificial intelligence when opening the event, "Cleverer than Humans? How Artificial Intelligence Changes Our Lives." Mr. Groth admitted that definitions were vague and that the level of knowledge in society left much to be desired. Generally speaking, he explained, AI referred to software that constantly self-optimized and self-developed through machine learning.

Nobody knows how a decision is made

In one example, Google fed a computer the rules for Go, a board game. The computer then taught itself the game – to the point where it beat the reigning world champion. Rule-based learning is the key principle in this context. The traditional computer only does what it was programmed in advance to do.

The flagship applications of artificial intelligence are autonomous driving – occasionally called ‘AI on wheels’ – and medical diagnostics, such as the analysis of bloodwork or X-ray images that may show diseases. In both areas, life and death can be at stake. It’s a disturbing fact that nobody knows exactly how the machines reach a decision or conclusion. Mr. Groth regards this as problematic because "people must have a choice."

A gut decision may also be good

For Mr. Groth, such questions are by no means academic in nature. In his book, "Solomon’s Code – Humanity in a World of Thinking Machines," he describes how he and his wife encountered such a situation in the early stage of her pregnancy: Based on a breast cancer diagnosis, doctors advised her to abort the pregnancy and start chemotherapy. The Groths decided against it.

Meanwhile, they have even had a second child. "Sometimes, trusting your gut is just the better choice," the AI expert writes in his book. Of course, he sees the risk of, for instance, a health insurance company insisting that a recommendation made by artificial intelligence is followed. Mr. Groth is adamant that this must not happen. Not wanting to leave these fundamental questions to the market, or more specifically, to American and Chinese Internet giants, he therefore proposes a global set of rules, a social contract that sets limits to what the "digital barons" can do. Mr. Geldner responded skeptically, given that current political developments around the world were in fact headed away from multilateralism. Mr. Groth admitted that resistance was to be expected, but remained true to his position. "It will take five to ten years," he said, "and we should start at the local and regional levels."

Americans still clearly in the lead

China is intent on becoming the global leader in artificial intelligence by 2030. As Mr. Groth sees it, however, the Americans are still far ahead when it comes to AI. And what about Germany? In terms of research, he sees the country among the international leaders, operating on world-class level. Nevertheless, in real-life application, he, like many other experts, sees plenty of shortcomings.

At the event, guests including high school students Rebekka Lederer, Timo Weiss, and Till Hermann had the opportunity to ask questions. Ms. Lederer, for instance, wanted to know the likelihood that moral questions would come to bear and not just the hoped-for benefits of artificial intelligence. Again, Mr. Groth was far from pessimistic, citing as an example a new role that was becoming established at American corporations: the ethics officer. Their task, according to Mr. Groth, is to ensure in working with IT specialists that ethical standards are adhered to.