It’s foolish to indulge in AI doom and gloom

General Purpose Technologies, so-called, change the world and how we all live. Most people agree that the new, potentially most powerful and transformative version – ranking with the printing press, electricity and the internet – is artificial intelligence (AI). But it is still a matter of debate whether machine learning will lead us more towards a utopia in which the quality of life will be permanently improved, or more to a

dystopia in which machines might not just out-fox us, but also destroy life as we know it.

Technophobia, which has stalked us for centuries, is nothing new. It is worrying, though, when technopioneers become techno-sceptics, themselves, as some have done recently, calling for the curtailment of AI research and development, or even going further to warn of sinister times ahead. They are right to emphasise that we do not as yet have a good understanding of the limitations and boundaries of machine learning. But hasn’t this always been the predicament of mankind?

Robert Oppenheimer, scientific director of the Manhattan Project, which was the US quest to develop nuclear weapons, felt personally liable for the terrifying destructive power of what he and colleagues had helped to create, and yet we have learnt somehow to live with his legacy. Even if today’s AI concerns are also especially trenchant, it is surely not possible now to stop what has started, and pretend that our curiosity and quest for knowledge can be corked.

We should think positively about AI’s potential. Anything that relies on or embeds copious amounts of data is fertile ground for this technology to explore and evolve better outcomes. Think for example of healthcare, including staffing solutions, patient care and monitoring in the community, as well as the use of more robust MRI and CT scans, injections and interventions to master and prevent non-communicable diseases such as cancer and diabetes. We haven’t even begun to think what the world in 2050 might look like if these ailments have been pretty much put out to grass.

More accessible, affordable, and personalised education, and life-long learning programmes are also likely to be interesting areas to which to apply AI tools and methods. Energy planning and optimal usage will benefit from AI. Managing and planning urban agglomerations, the environment and climate change mitigation are all going to use AI to a greater extent to realise important social goals. Consumer goods and services, transportation, entertainment, and financial planning and literacy will also ride along.

Yes, AI will cost jobs, especially for those involved in menial or repetitive clerical work, but also in higher skill and better paying jobs that may previously have escaped the reach of automation. But there is no reason to think that it will be the first technology ever to be a terminal destroyer of jobs. We never know where new technology-related jobs will be, or what demands they will spawn for new goods and services and workers equipped to deliver them. Just think of autonomous vehicles, and the multitude of jobs needed to build the accompanying infrastructure.

In any case, since our time is being defined as one of strategic competition with China where the goal is to establish the dominance of standards and values, we cannot afford not to make AI research a top priority. Defining and framing it in a way we think appropriate should be a leading concern. Government, AI tech firms and our institutions should make an urgent date to agree how to register and regulate AI research and applications in full transparency so we can all understand what we must embrace.

George Magnus is a research associate at Oxford University’s China Centre