Guard rails are needed around proliferation of AI

Scientific progress often comes with its moral quandaries. So it is with artificial intelligence, which promises to revolutionise our lives and boost the economy. The rapid proliferation of new technology that can write, talk and even see for us also presents us with the responsibility to negotiate some minefields.

AI simulates human intelligence. Its algorithms are used to think for us and help us achieve our goals – on building sites, in factories, hospitals, workplaces, classrooms and at home. It can predict our thoughts, with frightening accuracy. Not only can it help us write an essay, it can read and drive for us. Finance, education, transport, building, engineering, healthcare and the legal profession are among the industries it can transform.

As we report today, AI offers us many exciting applications in medical diagnostics. It can read radiology reports and help reduce the enormous workload for radiologists who are prone to getting tired and losing focus. While some practitioners welcome the extra help, some warn that too quick an embrace of the new technology may come at a big cost to the doctor/patient relationship.

Public health expert Professor Stephen Leeder said pressures on clinicians to rely increasingly on laboratory investigations for diagnosis rather than a detailed personal history from a patient had laid a foundation for AI to continue diminishing the need for human contact in making a diagnosis. He says it is becoming so depersonalised, he wonders where the doctor fits in.

The federal government has acknowledged the risks of harm associated with the new technology. It has released two reports that lay out the groundwork for regulation. Industry and Science Minister Ed Husic says Labor, like the rest of the world, is grappling with how to use AI safely and responsibly to balance its benefits with any risks.

An Industry Department discussion paper outlines the benefits AI could bring in building and engineering, legal services and storage of hospital patient data. Consulting firm McKinsey estimates AI could add up to $4 trillion to the Australian economy by the early 2030s.

Loading

The discussion paper also warns of risks associated with generative AI such as ChatGPT. Its creator, OpenAI chief executive Sam Altman, has spruiked the potential for chatbots to improve our lives. But he has also admitted tech companies are in danger of unleashing a rogue artificial intelligence that could cause harm if governments fail to intervene.

“If this technology goes wrong it can go quite wrong, we want to work with the government to prevent that from happening,” he says.

Concerns have been raised about the potential for harmful systems that challenge human notions of justice and morality in fields such as government services, banking and military decision-making. Some doomsayers have gone as far as predicting that programmers could accidentally create a superintelligence that could – either deliberately or accidentally – wipe out humanity.