Jack Ma, head of Chinese internet retail giant Alibaba, is the latest tech visionary to suggest AI will harm society. He’s .
His warnings follow those of luminaries including Stephen Hawking, Bill Gates and Elon Musk, for AI research to focus on beneficial applications and to be made robust against runaway failures that shut down electricity grids or even lead to some kind of .
So you could forgive the British public, fed a diet of such high-profile comments, dark sci-fi visions and , for feeling negative about AI and machine learning, the computing method behind most AI. However, attitudes are relatively balanced so far, according to an opinion poll for the Royal Society released this week.
Advertisement
It found broad support for the potential use of this technology – even if there are some common concerns, such as losing jobs to robots.
Ipsos MORI, which carried out the poll for the Royal Society, found that people expect AI’s algorithms to be more honest, quashing errors made by tired and emotional humans in deciding who gets a mortgage, insurance or a loan, for instance.
And, if due regard is given to privacy and security of medical data, most of those quizzed hope healthcare driven by machine learning will be more accurate at detecting disease in medical images and make better diagnoses earlier.
Areas of concern
It wasn’t all positive. Faring less well were self-driving cars, one of the most visible tests of machine learning, being essentially robots we sit inside. More people saw downsides than upsides of this application at the moment.
And one broad streak of pessimism, much discussed of late, came through: people are concerned about their ability to keep their jobs as intelligent machines inch ever closer to reality. “The potentially negative consequences of machine learning for employment were a repeated concern,” said Ipsos MORI in a commentary on the it conducted.
Other concerns were less obvious. Fears were raised that ever smarter systems could cause treasured human skills to wither on the vine – like the ability to drive or to read maps. Some dismissed the very premise of autonomous cars, saying they regarded driving as fun, and that they do not want to surrender it to intelligent machines.
The poll has been used to inform , also launched this week, on what needs to be done to ensure this field thrives.
One chief conclusion is that the AI industry needs public opinion on its side, and so the Royal Society recommends developers spend at least some of their research cash on meaningful engagement. That’s sound advice.
We can already see the consequences of a lack of engagement where driverless vehicles are concerned: while research and testing on autonomous cars presses ahead, no one seems to be consulting those who drive for a living over what will happen to their jobs. That is simply not good enough.
Developers of AI – and those who fund them – must be transparent and open about their aims, concerned not only with technical progress but with their technology’s potential for profound social impact. Only then can the power of machine intelligence be made to benefit everybody.
Read more: Instant Expert: Artificial intelligence
