Dr. Frank Marcinkowski is a professor at the Institute of Social Sciences at Heinrich Heine University Düsseldorf and a member of the Artificial Intelligence Opinion Monitor, a research organization devoted to gauging views and opinions regarding artificial intelligence voiced in public discourse.
The AI Observatory and the MeMo:KI research team engage in regular dialogue and conduct joint projects and surveys like the special survey “AI at Work.” Data gathered by MeMo:KI will flow into the AI metrics and indicators developed by the AI Observatory.
1. What is your research project about, both generally and with regard to AI at Work as special topic?
The function of MeMo:KI, also known as the Artificial Intelligence Opinion Monitor, is to provide ongoing, long-term monitoring and gauging of German public opinion regarding Artificial Intelligence (AI). As part of this mission, we document public debate on the subject in both traditional and social media, seeking also to identify relevant intersection points. We are interested in general public attitudes to AI as a topic, and also in how specific issues are navigated such as use of AI in the COVID-19 pandemic response, arising discrimination and ethical issues connected with AI and the impact of AI on employment.
2. What questions is the project aimed at answering?
There’s a noteworthy contradiction behind our interest in this data, as on the one hand the German government and EU have promised that Europe will take its own independent and human-centric approach to AI. At the same time however, the vast majority of people here exhibit a remarkable disinterest in the topic and are vaguely informed at best regarding AI. As a result, surprisingly little conflict has been stirred by the technological changes surrounding the advent of AI in comparison to other technological breakthroughs at the time of their introduction (such as atomic energy, biotechnology and genetic engineering). For governments this is a comfortable situation, taking pressure off them to act. But such passivity regarding this technological paradigm shift is undesirable if the goal is for broader society to play a role as educated participant in the changes affecting it. Thus our simultaneous objective is exploring ways of furthering popular “AI literacy” and stimulating interest among the citizenry in the far-reaching societal implications of this change.
3. What are some of the initial project findings?
Having conducted public opinion surveys on AI in short intervals since May 2020, we have observed nearly no change in awareness of AI or in opinions on specific applications used in various areas. This is undoubtedly in part due to COVID-19 taking center stage, pushing every other topic out of people’s minds, more or less. Roughly half of survey participants expressed a low level of interest in AI, while roughly one in five say they are very interested. Two-thirds of survey participants are in favor of the use of AI in industrial production, which is a significantly greater approval level than for any other area of application (such as just over 40% regarding traffic and less than 10% in politics and the courts).
Our surveys have not indicated significant anxiety over job losses due to AI. Currently only 6% of respondents say they fear losing their job to intelligent machines. Roughly 20% do believe that the widespread deployment of AI will result in mass unemployment. The dominant view is that AI has little potential for altering work and employment over the medium term. Expected changes concern improved occupational safety and health as well as decreased social interaction at the workplace.
4. Where can I find out more about the project?
MeMo:KI findings are published online – there’s a dashboard available on our Project page for retrieving data from our ongoing observation work as well as factsheets on special topics we have studied. The factsheet on AI at Work can be found there, for example. And my staff and I are happy to answer questions via e-mail as well (memoki[at]hhu[dot]de).
5. What innovative AI application would you personally most like to see?
In my day-to-day work as university professor I am always amazed at the shifts in student demand and registration for study places and places in individual courses. Some predictive analytics that could shed a little light here would be desirable, making course planning a lot easier. I believe there are a lot of possibilities for useful AI applications in a university setting, as long as these are developed and implemented in a manner compatible with social responsibilities.