The reason people talk about “artificial intelligence” today is simple: marketing. Computer and cognitive scientist John McCarthy created the term “because we had to do something when we were trying to get money for a summer workshop in 1956,” he once said. AI sounded new and exciting and like it was worth some funding. Almost 70 years later, AI is everywhere. But what kind of intelligence is meant here?
Private lecturer Rudolf Seising found answers to this question on Monday evening at the Munich Economic Debates hosted by the Ifo Institute and Southgerman newspaper. He leads a project that researches the history of AI in Germany. “What can we learn from the development of AI for the future?” was the title of his lecture.
What ennobles AI as “intelligent” is historically shaped by mathematics and computer science, says Seising. As early as 1950, the British mathematician Alan Turing asked “Can machines think?” without knowing the answer at the time. Generations of researchers worked to translate language into codes and capture the human mind in logical formulas and algorithms to build such “thinking” machines. This legacy can still be seen in artificial intelligence today: it is a narrow interpretation of intelligence, a non-human and calculating variant.
“We judge people, so we should also think humanely”
History can reveal a lot about the present with AI – and possibly also about the future. Seising’s actual area of expertise is the past, he emphasizes. However, he sees that since the release of the Chat-GPT chatbot, more and more scientists are jumping on the AI bandwagon. This could narrow the field of questions asked. Because if you have a hammer, you only see nails. And anyone who has an AI may always see problems as being solvable with AI.
One field that would be ideal for this could possibly be law, was one suggestion from the audience. Shouldn’t judges just calculate facts? Definitely not, the otherwise calm Seising answers a little more vehemently for the first time and sits up straight: “We judge people, so we should also think humanely.”
Seising is not really worried about the future of AI. He may know the history of the technology well enough to correctly assess its strengths and weaknesses. But whether others value such nuances seems uncertain. Because while huge amounts of money flow into AI development, he cannot find any funding for his research. “If anyone would like to give me one to two million, please feel free to get in touch,” he grins at the audience. That means: If all the money flows into technology and none into history, there may soon be very powerful AIs – and no one left who can classify them.