Future AI models should be based on human values, says QF professor

With technology being innovated and improved at a rapid pace, artificial intelligence is creating a huge qualitative shift in the way humans accomplish tasks – what would once take hours now takes minutes, and is done with minimal effort.

“Artificial Intelligence (AI) is a conscious and systematic effort that human beings make to enhance their mental and physical abilities,” says Dr. Georgios Mikros, Professor in Master of Arts in Digital Humanities and Societies, at the College of Humanities and Social Sciences (CHSS), at Qatar Foundation’s Hamad Bin Khalifa University (HBKU).

“AI systems have been designed to solve a real problem for humans in various application domains. Their performance is directly related to the narrowness of their scope and the abstractness of the task. In well-defined closed problems, where vast data exist, AI systems deliver what they are designed to provide.”

However, as the application domain is becoming wider, the task is becoming increasingly general in scope, and the training data are sparser, then the AI systems become increasingly unable to achieve the required level of performance, according to Dr. Mikros.

He believes AI systems should be engaged in real-life applications only when tested thoroughly in large-scale, highly diversified experimental conditions. Their performance should be sufficient and stable even when data are sparse and fragmented.

Dr. Mikros notes that humanities must be seen as the cover of this technological ecosystem. “Humanities is the bridge that links the new digital era to the core values of our human civilization. Skills like critical thinking, communication, empathy, mindfulness, resiliency, decision-making, and self-awareness should be made the foundation of a new ethos that will make our societies technologically advanced and, at the same time, profoundly humane and sustainable.

“Humanities is the key to keep AI and other disruptive technologies within the bounds of the human domain. They are vital to developing the scientific and technological foundations for AI that are beneficial to humans and humanity according to ethical, social, and cultural values.”

With a strong belief that humanities should be part of the foundation of AI, Dr. Mikros underlines that humanities should provide the roadmap for the future development of AI, enabling technologies for systems that seamlessly fit with complicated social settings and dynamically adapt to changes in our environment to empower people.

He explains: “Artificial intelligence is a cover term that involves many different types of technologies that aim to create machines that can exhibit intelligent behavior. Like every form of intelligent behavior, AI is based on the developmental nature of learning.

“We "teach" machines to "learn" by using the analogies met in our physical world. We expose them to large amounts of data, and we let them discover meaningful patterns. Computers can find associations inside datasets by using algorithms that approximate natural human learning processes.”

Dr. Mikros noted that the human mind offers highly abstract and complex cognitive functions that are hard to replicate artificially. For an AI system to simulate the human mind, it should at least develop a causal model of the world that supports explanation and understanding, rather than merely solving pattern recognition problems, as well as grasp fundamental and intuitive concepts of physics and psychology and learning-to-learn in order to acquire and generalize knowledge to new tasks and situations rapidly.

“Human intelligence is the product of nine million years of evolution of the brain and its primary communication tool – the human species' language. It is a refined intellectual capability that allows us to think, learn, and adapt to different situational contexts, grasp complex concepts, reason, solve problems, make judgments, express feelings, and emotions, and communicate with other humans.”

AI can offer advanced analytical capabilities in narrow fields of inquiry and surpass humans in several cognitive tasks, including information retrieval, gaming, pattern association, and image recognition. However, human intelligence's uniqueness is based on abstract emotions like self-awareness, passion, and motivation that enable humans to accomplish complex cognitive tasks with ease and learn new skills with limited exposition to new training data, according to Dr. Mikros.

He indicated that the effects of AI in our societies are already profound. “In the near future, we expect this change to be generalized. Nearly every aspect of our social structure will be transformed and adapted to the new digital environment. Our workplaces will be improved with augmented human capabilities.”

For example, when AI takes over repetitive or dangerous tasks, it frees up the human workforce to do work, leaving them space for working, on tasks that involve creativity and empathy, among others.

If people are doing work that is more engaging for them, it could increase happiness and job satisfaction, says Dr. Mikros. Healthcare will also dramatically improve, reducing costs and increasing efficiency, leading to better patient care. Individualized therapies and cutting-edge diagnostics will revolutionize treatment options leading to better health outcomes for the general public.

He says: “Moreover, autonomous transportation will lead to improved traffic flow, saving hours of lost productivity, and improving all commuters' quality of life. Freed up from stressful commutes, humans will be able to spend their time in other ways. The battle against crime will be made easier as facial recognition and other crime preventive AI-enhanced technologies will be massively available. Moreover, our justice procedures will become fairer as AI will help judges to apply justice in an unbiased way. AI under the proper regulatory framework can make our societies more cohesive, reduce social conflicts, and strengthen our community bonds.”

AI means different things to each of us, but everyone can be sure that the progress obtained cannot be undone. AI changed and will continue to change our virtual and physical environments in sometimes unexpected ways. However, “the core values of humanities should be the base on which all further AI progress should be established. We should learn to adapt and transform AI to serve the broader interests of humanity,” Dr. Mikros says.

“Our collective responsibility is to realize that humans cannot be replaced by technology; we must ensure precisely how these two actors will intersect and intertwine in our future world.”