AI could lift productivity and living standards, but without stronger government leadership, workers risk missing out on the benefits at the same time as bearing the costs.
Part 1 yesterday discussed how AI could restore productivity growth and living standards, but the challenge would be to ensure that the benefits were fairly shared. Today, Part 2 discusses these distributional risks, and the adequacy of the Government’s AI strategy.
Yesterday we concluded that the Government faces a dual challenge. It wants to facilitate the adoption of AI so as to lift future economic growth and relieve cost-of-living pressures, while at the same time avoiding mass unemployment and ensuring a fair distribution of the benefits from AI.
As Nobel laureates, Daron Acemoglu and Simon Johnson say, if we want to avoid a major fallout of displaced workers and consequent hardship for them at least, then we must ensure an institutional set-up that induces “a direction of technology that creates new tasks and jobs for workers of all skill levels and an institutional framework that enables workers to share productivity increases with employers and managers.”
The Government’s strategy
About a month ago, the Prime Minister, Anthony Albanese, outlined the Government’s strategy for encouraging the take-up of AI in Australia’s interests. In addition, the ministers most directly responsible, Tim Ayers and Andrew Charlton, released a National AI Plan.
However, the emphasis in both Albanese’s speech and in the AI Plan is mainly on how to promote the take-up of AI, while minimising energy and environmental costs to the community, and minimising other risks, including bias, privacy breaches, disinformation and cyber threats, as well as protecting the rights of authors and other artists.
In fairness, the AI Plan does have a section, “Support and train Australians”. However, this section is also mainly about fostering the development of AI with the Government promising to “take early action to build a workforce that can thrive in an AI-enabled economy, with initiatives that are underway to boost digital skills, expand training access, and grow an inclusive pipeline of AI-ready workers.”
But, while AI can boost productivity, it also poses risks to workers. To be effective in practice, deploying these technologies should involve meaningful consultation with workers, including in the design of AI systems. In principle, I think the Government recognises the need to consult. There is, however, no apparent recognition so far by the Government of the need for a strategy to steer AI toward complementing humans rather than imitating and replacing them.
As a group of more than 200 international economists, including 16 Nobel Prize winners, have recently declared, “AI complements humans when it enlarges the set of things worth doing, broadens access by lowering costs, and enables people to do what was previously out of reach.” In contrast, AI substitutes for humans when it simply transfers certain tasks from workers to models. The first path expands the economic pie, while the second path merely redistributes value.
But as Erik Brynjolfsson, one of the organisers of the above declaration has said, “Designing AI to imitate humans pushes technology toward substitution rather than augmenting, concentrates power in the hands of those who own the systems, and leaves ordinary workers with fewer opportunities to create value.”
Furthermore, we know that AI has the potential to augment services, rather than simply lowering their cost. Indeed, AI has the potential to reshape the economy even more than previous waves of automation because, rather than merely automating individual tasks, it lowers the cost of cognition itself.
Thus, AI can complement people’s judgement when it increases the value of judgement, tastes, accountability, relationships, and other human capabilities that AI does not provide. For example, AI can complement the skills of professionals such as lawyers, doctors and scientists, allowing them to increase their services, where demand has long been constrained by the scarcity of supply of these skilled professionals.
But such favourable outcomes are not inevitable, and that is why so many eminent economists signed the declaration cited above. Governments therefore have a role to play and they need to get on the front foot.
In fairness, the Ministers do say in their foreword to the AI Plan that “Success will rightly be measured by how widely the benefits of AI are shared, how inequalities are reduced, how workers and workforces can be supported, and how workplace rights are protected. AI should enable workers’ talents, not replace them”.
But the critical question is how to achieve an equitable outcome. For their part, the Ministers seem to put all their faith in their commitment to “a consultative approach to AI adoption in the workplace, and [how they] will bring together government, unions and business, on issues including the uplifting of AI skills and training of all Australians.” Similarly, the Plan also says, “Workers’ voices and union engagement must guide decisions on technology adoption to ensure fairness and protect rights”.
Unfortunately, however, I would question whether this admirable commitment to consultation is enough. I fear that if we rely mainly on employers to lead the way, they will focus on reducing costs, and too often ignore how services can be enhanced. Replacing a worker is easy, while changing production processes to enable people to accomplish more things is harder, takes longer, and the benefits are more widely spread.
Thus, if we want to successfully maximise the opportunities from AI, the Government needs to get on the front foot in identifying those opportunities, promoting them and even helping to develop and implement them. At the very least, governments should be examining how to improve the provision of public services, such as nurses consulting AI in interpreting symptoms and how teachers could use AI to reduce their administrative load and personalise instruction for each individual student.
But at this stage, at least, our Government is providing very little leadership and seems inclined to leave all the initiative and even much of the regulatory system to business and the market.
Read Part one:
Michael Keating is a former Secretary of the Departments of Prime Minister and Cabinet, Finance and Employment, and Industrial Relations. He is presently a visiting fellow at the Australian National University.