*Article by Maria Kouzinopoulou
Artificial intelligence is in every aspect of our lives today, from mobile phones and smartwatches, movies and virtual assistants to medical diagnostic devices and classrooms. At the same time, it is estimated that by 2030 the artificial intelligence market will contribute $15.7 trillion to global economic activity. So the impact of artificial intelligence is and will continue to be huge, but what are the values that should govern the use of this technology? Scientists of NCSR Demokritostried to give an answer to this question through research.
Researchers at the Institute of Informatics and Telecommunications (NCSR Demokritos) conducted a meta-analysis of major studies from around the world, which compiled guidelines for the ethical use of artificial intelligence. Specifically, ten studies, published from 2018 to 2023, were investigated, presenting tens or hundreds of scientific articles and policy texts with guidelines on the ethical use of artificial intelligence, published by private and public agencies, civil society organisations and academic institutions. The results of this analysis, published in the proceedings of this year’s international conference on the ethical and social impact of IT “ETHICOMP”, were presented by the researcher and one of the authors of the study, Xenia Ziouvelou, during the Summer School on Artificial Intelligence, organised by the “Hellenic Institute of Advanced Studies” (HIAS) in NCSR Demokritos.
According to the results of this meta-analysis, despite the fact that growing concerns raised by the rapid development of artificial intelligence systems have led to dozens of guidelines for dealing with ethical issues, the studies find broad convergence on four dominant ethical values that should govern the creation and use of artificial intelligence systems: Accountability, privacy, fairness, transparency.
However, there are variations in the definition, interpretation and scope of these values, and several of the guidelines presented by the studies are too abstract creating difficulty in adopting them when designing artificial intelligence systems. In addition, it was found that the proposed guidelines focus mainly on algorithmic design and decision-making processes, which however is only one part of the ethical design of artificial intelligence systems, without highlighting the human factor involved in the creation and use of these systems.
The researchers emphasise in their study that in order to have a holistic approach to the guidelines in the ethics of artificial intelligence, an understanding of human values and the role of humans creating the technology is necessary. In this context, the “AI Politeia Lab” laboratory of the Institute of Informatics and Telecommunications at NCSR Demokritos, headed by Mrs. Ziouvelou, is designing a software application, where through educational material it will aim to help researchers and programmers develop themselves and their system while creating it.
Summer School on AI: a field of dialogue
The Summer School on Artificial Intelligence, organised by the “Hellenic Institute of Advanced Studies” (HIAS) in NCSR Demokritos, founded by Greek scientists of the diaspora, was a unique meeting ground for Greek experts in artificial intelligence from universities and research centers of the diaspora and Greece.
During the three days of its implementation, approximately 450 undergraduate, postgraduate and doctoral students, as well as researchers, attended 22 lectures on artificial intelligence. The speakers included, among others, Greek scientists from important educational institutions around the world: Harvard, Columbia, ETH Zurich, University of Texas at Austin, University of Toronto, Rice University, University of Wisconsin and University of Pennsylvania.
A large portion of the presentations were related to the latest developments in artificial intelligence systems. Also, the presentation of the uses of artificial intelligence in a number of sectors was an important aspect of this year’s event, although this section was not completed due to a fire in Ymittos. Just before this break, Columbia University Biomedical Engineering and Radiology Professor Elisa Konofagou presented the use of artificial intelligence in studying the results of imaging tests in a range of diseases: Cancer, Alzheimer’s and heart disease. Among other things, artificial intelligence has contributed to the prognosis of the progression of breast cancer, the identification of genetic factors for the onset of Alzheimer’s disease and the detection of coronary heart disease. In fact, as he said, the operation of the “Center for Innovation in Imaging Biomarkers and Integrated Diagnostics”, an interdisciplinary center, is being launched at Columbia University in the fall of 2024, in which, through the widespread use of artificial intelligence and Data Science, the development of next-generation assays that will play a key role in the prevention and treatment of diseases will be pursued. The Director of the Center will be Despina Kontos, who is also of Greek origin, a professor of Radiology and an IT scientist specialising in artificial intelligence, while the deputy director will be Eliza Konofagou.
Artificial intelligence researcher at Google DeepMind, Yannis Assaël, presented the possibilities of using artificial intelligence in the restoration of ancient Greek inscriptions through the Ithaca model, which, as he revealed, is already being tested on Latin inscriptions, while a group of his colleagues is experimenting with ancient Chinese inscriptions.
Members of the research team of the Speech Processing Institute at the Athena research center, led by the director, Vassilis Katsouros, presented Meltemi, the first Greek Large Linguistic Model of Artificial Intelligence, and the challenges it “hides”, as, as he described, the training of a language model in Greek is complex and needs significant resources.
Opening the proceedings of the Summer School, the president of HIAS and chair of the Department of Applied Mathematics at Harvard University’s School of Engineering and Applied Sciences, Petros Koumoutsakos, reviewed the history of Information Technology and artificial intelligence and then, talking about the future , emphasised that “computers are here to stay and we need to think about ethics. What I find fantastic about artificial intelligence is that if you think about all the great philosophers and scientists in ancient times, they could “cross over” different scientific views. At some point the different scientific views got separated, and I think artificial intelligence is bringing scientists of different disciplines together again.” “We need to understand artificial intelligence and not see it as a monster or an enemy that threatens human existence,” he added.
The Director of Institute of Informatics and Telecommunications at NCSR Demokritos, Vangelis Karkaletsis, also underlined the need for artificial intelligence research to be interdisciplinary, involving researchers from other sciences.
The actions of HIAS concluded on 4th July with a conference on Sustainability, at the Eugenide Foundation.
Source: amna.gr
Article by Maria Kouzinopoulou