Public Not Convinced On The Benefits Of AI

Despite the progress made with AI in recent years, for many, it remains as much a myth and a mystery as a cold, hard factor in their lives.  This perhaps goes some way towards explaining the continued hesitancy towards the technology that was revealed in the Global Attitudes Report from Oxford University’s Internet Institute.

The report reveals a public that is broadly speaking divided down the middle on the impact of AI in public life, with those in the East seemingly far more comfortable than those of us in the West.

The research was based upon the data generated by the 2019 World Risk Poll produced by the Lloyds Register Foundation.  Their survey examines the public perception of global risks, with 142 countries participating.

Perceived risk

“Understanding public confidence in AI and machine learning is vital to the successful implementation of such systems in government,” the researchers explain.

The researchers were particularly keen to understand how people feel AI will contribute (or not) to society in the next twenty years.  The results highlight clear regional differences, with clear skepticism present in North America and Latin America, where at least 40% believe that AI will be harmful.

This compares to South East Asia, where just 25% believe this will be the case, and to East Asia, where just 11% have such concerns about the future development of AI.  This was most pronounced in China, where just 9% of people thought AI would be a risk to society.

There was also a visible sectoral divide in the responses, with those in business and government most optimistic about the potential for AI to benefit society.  On the flip side, those in construction and service workers were least confident about the benefits.

“Our analysis suggests that putting AI to work for good governance will be a two-fold challenge,” the researchers conclude.  “Involving AI and machine learning systems in public administration is going to require inclusive design, informed procurement, purposeful implementation and persistent accountability.  Additionally, it will require convincing citizens in many countries around the world that the benefits of using AI in public agencies outweighs the risks.”

Facebooktwitterredditpinterestlinkedinmail