Article from the AFR reporting on the ASIC Annual Forum yesterday and the focus on AI from a variety of people, including the Governor of the Reserve Bank, the Productivity Commission Chair, the ASIC Chair … and, if you read right to the very end of the article, me 😃 It seems my autopilot story hit a nerve.
While this is a potential issue I identified with AI, I did focus most of my comments on the productivity improvements we will see from AI as a powerful tool for business. Also that we now have generative AI inside our businesses by stealth as a result of our existing tech providers (like Microsoft, Salesforce etc) providing AI copilots as part of their existing solutions. I love the idea of an AI copilot reading all my emails and preparing amazingly accurate draft responses to them – accurate because the AI knows how I write as it has read every email I have ever sent.
If interested, here is my bit from the AFR piece (as it is behind a paywall).
“Norton Rose technology lawyer Nick Abrahams warned that AI and automation could lead to key skills being lost, sometimes with catastrophic results. He cited the example of a 2009 Air France flight which crashed after a pilot had to take over from an autopilot, but had over time lost the skill to fly the plane safely.
“We need to be careful with generative AI, that we don’t become deskilled with what it’s providing to us and that we maintain the competence and the understanding of how it’s making decisions, what parameters those decisions are being made under, and that we can effectively be able to explain how the AI came to what it came to.”
Full Article from the AFR
Artificial intelligence could help repair Australia’s lagging innovation record, unleash waves of productivity gains and lift the output of workers such as software coders and call centre operators, new productivity commissioner Danielle Wood says.
Reserve Bank governor Michele Bullock also said AI could improve productivity, and used a personal example to suggest how.
AI could help drive more efficient R&D, lifting innovation and fuelling successive waves of productivity, new Productivity Commission chief Danielle Wood says.
“I’m optimistic on this … I think potentially AI has a very big impetus for productivity improvements,” Ms Bullock told an ASIC forum in Melbourne on Tuesday.
“I have a daughter who’s a nurse, for example. Are there things that AI could assist with to allow her more time to be actually caring for patients rather than filling in forms or doing things like that?” she said.
“There’s plenty of opportunity here to think how can AI help to take on some of these tasks. That gives people opportunities to do things that actually are of more value.”
They were speaking at the Australian Securities and Investment Commission’s annual summit in Melbourne on Tuesday.
Productivity Commission chairwoman Ms Wood told the forum AI could affect as much as 80 per cent of the economy.
“There’s some pretty phenomenal research coming out about the way it impacts productivity in jobs that are there now,” she said.
She cited examples of coders doubling their output and call centres increasing output by 15 per cent by using AI.
“AI itself can be a driver of innovation. People can use it to do better R&D, to improve the way technology is used, and that is really where you can get big unleashing productivity gains.
“This has the potential to be one of those technological changes that unleash successive waves of productivity gains. It comes with challenges and risks … but it’s hard not to be optimistic about the potential.”
ASIC chief Joe Longo agreed the promise of AI was huge, but said regulators were still getting a grip on what they are dealing with, and it was important to consider the impact on its users and investors.
“I think we need to be a bit sober about how we approach this. From a regulatory perspective … there are some very profound issues that need to be discussed.
“AI is data-driven and there’s a big question about the quality of the data that’s feeding into these models,” he said.
“The big issue has been explainability. How can we explain what the algorithm is doing, whatever it is the AI is doing? In the end, I would like that to be explainable by human beings, not by another algorithm.”
”And there’s a very day-to-day types of issues, insurance premiums, lending decisions – and there’s a lot of work that’s been done on financial stability and the potential for AI-driven systems to be disrupted with markets.”
Sydney University expert Nick Davis said companies needed to do an inventory of their usage of AI.
“Where are we right now in terms of AI? What form does it take in your system? Most organisations cannot answer that question at all right now,” Professor Davis told the ASIC forum.
“I talked to about 300 different organisations for a big project on the corporate governance of AI. And I think it was one organisation that had done in an inventory out of the 300.”
This needed to be married to a governance system and for directors to become skilled in understanding the technology, he said.
Norton Rose technology lawyer Nick Abrahams warned that AI and automation could lead to key skills being lost, sometimes with catastrophic results. He cited the example of a 2009 Air France flight which crashed after a pilot had to take over from an autopilot, but had over time lost the skill to fly the plane safely.
“We need to be careful of with generative AI, that we don’t become deskilled with what it’s providing to us and that we maintain the competence and the understanding of how it’s making decisions, what parameters those decisions are being made under, and that we can effectively be able to explain how the AI came to what it came to.”