In his latest book, George Zarkadakis (PhD, Artificial Intelligence in Medicine, 1990) questions the current state of our democracy and economy, and how AI can either help to reform it to benefit our voices as citizens or assist in our downfall.
Although three-time City alumnus George Zarkadakis trained as an engineer and works as Digital Lead at Willis Towers Watson, he considers himself to have a firm grounding in the humanities. It’s a connection he finds more important now than ever.
“Technological innovation, marvellous as it is, has reached a point where we cannot simply keep pushing the boundaries while disregarding the broader implications on society. In my own field of artificial intelligence this realization is rather profound”, George says.
According to him, AI algorithms already have a significant impact on how we govern ourselves in public as well as in the private sphere.
“Understanding that impact is imperative in order to design and develop systems that uphold our moral values, advance human wellbeing, and support our freedoms and liberties.”
In his latest book Cyber Republic – Reinventing democracy in the age of intelligent machines, George sets out to reinvent liberal democracy and the economy, both of which he considers to be under threat.
“Citizens are losing trust in the liberal system of government for two main reasons,” he explains. “Firstly, they know that powerful economic lobbies yield asymmetric power in influencing policy making. Secondly, given the huge and constantly rising economic and social inequalities, citizens also realize that the social contract of liberal democracy is broken, and that their livelihoods and incomes are under threat due to the digitalization of the economy and the automation for work by AI systems.”
George believes that to avoid authoritarianism, we must make politics more transparent and participatory, and rethink the digital economy so that its economic value is distributed more equitably.
“In my book I am questioning the dominant dogma of AI, which is system autonomy. I would like to argue that most of the ethical dilemmas and problems with AI systems today stem from this dogma,” he says.
“I therefore suggest that we should rethink AI systems as elements in broader cybernetic systems where humans remain in the loop and in control. An example of such “cybernetic AI” systems would be intelligent conversational agents that help citizens arrive at consensus in citizen assemblies.”
The thought of an increasing amount of AI in our everyday lives can seem scary, and George doesn’t believe this is an unfounded fear.
“You are right to feel that AI is a threat to jobs, privacy and safety. Politicians, ethicists, economists, regulators and others are struggling to develop frameworks to reduce the threat of AI. I think that they will ultimately fail because, once you accept AI autonomy as a given, there are only two alternatives: either AI will advance itself beyond the control of regulatory frameworks, or humans will push back and demand that AI systems are not allowed to advance”, he says.
Neither of these options would benefit human civilisation, according to George, so AI needs to be harnessed by humans.
“As I mentioned before, it must be rethought as software systems embedded in human systems, rather than as autonomous systems in competition with humans, as they currently are.”
One example of where AI has already been of great benefit to our lives is health and medicine, a subject George researched in his PhD at City.
“Over the past few years, we have seen the application of deep networks in solving problems in medical diagnosis, particularly in pattern recognition problems, that were impossible to solve algorithmically before. These advances require that we rethink how we train doctors and nurses, and how we organize our health systems.”
George stresses again the importance that AI doesn’t become an “add-on” technology but must be integrated into new designs for health and medical systems where the role of human doctors and nurses remains key.
With his in-depth knowledge of AI and how it’s becoming a feature we hardly even think about, does George use smart technology himself?
“Our daily lives are increasingly controlled by AI algorithms. They decide what we should read, what news we should consume, what products and services to buy, how to behave, how to work. I am very worried about the degree of influence that AI has on people’s behaviour and psychology, and try to avoid as much as possible my exposure to “everyday AI” in social media, etc. I prefer to use AI in solving business problems where I see a clear utility.”
What about the world outside of business? In addition to non-fiction, George Zarkadakis has written novels, poetry and plays. He says he simply wants to share the stories in his mind with others. Could AI genuinely replicate the creativity of human minds?
“In art there is a difference between a work of art and a masterpiece. I think that AI is already very close to delivering works of art, including novels and paintings and music, that are of high quality and commercial value”, he says.
And that leaves us with the masterpieces.
“This is a very difficult question to answer for two reasons. First, we need to agree what a masterpiece is. I would like to propose that a masterpiece is a work of art that reveals something new and unexpected to human consciousness. But the problem with that definition is that it requires an observer sensitive enough to appreciate the new and the unexpected. My worry is that, as culture is systematically dampened down across the world, it will be very difficult – if not impossible – for humans to appreciate a masterpiece in the future, whether created by a human or a machine.”
George Zarkadakis participated this March in the alumni event “Artificial Intelligence for Solving Real-life Industrial Problems“ as a member of the expert panel. If you would like to watch a recording of the panel discussion, please get in touch with the Alumni Team.