Olle Häggström: Large Language Models, AI Risk and AI Alignment

Research seminar with Olle Häggström, a professor of mathematical statistics at Chalmers University of Technology, an affiliated researcher at the Institute for Futures Studies, and a member of the Ro

European Sociological Review 31: 312-25. Abstract We study the biases that arise in estimates of social inequalities in children’s cognitive ability test scores due to (i) children’s misreporting of soci

Journal of Ethnic and Migration Studies, 41(3), 399-424. DOI:10.1080/1369183X.2014.920690 Abstract Despite the importance of adequately measuring ethnicity to keep track of ethnic disparities in importa

in: Artificial Intelligence Safety and Security (ed. Roman Yampolskiy), CRC Press. Abstract Developing a superintelligent AI might be very dangerous if it turns out to be unfriendly, in the sense of hav

Prof. Dr. Armin Schäfer, Institut Für Sozialwissenschaften, Universität Osnabrück ABSTRACTAs turnout has declined in many developed democracies, it has also become more unequal. Recent studies show tha

Venue: Institutet för framtidsstudier, Holländargatan 13 in Stockholm, and onlineResearch seminar with Olle Häggström, a professor of mathematical statistics at Chalmers University of Technology, an affi

Nature Scientific Reports Abstract When someone violates a social norm, others may think that some sanction would be appropriate. We examine how the experience of emotions like anger and disgust relate

in: Technology in Society. 63, 101396 Abstract:Artificial Intelligence (AI) in the form of Deep Learning (DL) technology has diffused in the consumer domain in a unique way as compared to previous gene, i.e., by being added to preexisting technologies that are already in use. We find that DL-algorithms for recommendations or ranking have been infused into all the 15 most popular mobile applications (apps) in the U.S. (as of May 2019). DL-infusion enables fast and vast diffusion. For example, when a DL-system was infused into YouTube, it almost immediately reached a third of the world's population. We argue that existing theories of innovation diffusion and adoption have limited relevance for DL-infusion, because it is a process that is driven by enterprises rather than individuals. We also discuss its social and ethical implications. First, consumers have a limited ability to detect and evaluate an infused technology. DL-infusion may thus help to explain why AI's presence in society has not been challenged by many. Second, the DL-providers are likely to face conflicts of interest, since consumer and supplier goals are not always aligned. Third, infusion is likely to be a particularly important diffusion process for DL-technologies as compared to other innovations, because they need large data sets to function well, which can be drawn from preexisting users. Related, it seems that larger technology companies comparatively benefit more from DL-infusion, because they already have many users. This suggests that the value drawn from DL is likely to follow a Matthew Effect of accumulated advantage online: many preexisting users provide a lot of behavioral data, which bring about better DL-driven features, which attract even more users, etc. Such a self-reinforcing process could limit the possibilities for new companies to compete. This way, the notion of DL-infusion may put light on the power shift that comes with the presence of AI in society.

Institutes, companies and organizations connected to futures studies or future aligned consultancy. The Institute for Futures Studies is not responsible for the content of these web sites and many of t

The more things change, the more they stay the same. A follow up of participants in Social Fund financed projects

Research report 2014/5, 77 p. Every year in Sweden, over one hundred thousand job-seekers are assigned to local labour market policy measures, of which a large proportion are financed with money from t