Ethics challenge for AI and IAG
Big city: Coming to grips with the ethical challenges of Artificial intelligence
Fresh from its monstering three months ago at the hands of the banking Royal Commission, the local insurance giant IAG has turned up as the primary sponsor of a new institute that aims to create ethically aware artificial intelligence.
IAG has joined Data61 and the University of Sydney as founding partners of the Gradient Institute, a not-for-profit that would “research the ethics of artificial intelligence and develop ethical AI-based systems that will provide better outcomes for individuals and society.”
The Gradient Institute will be housed inside the wholly-owned IAG subsidiary Ambiata, a data analytics company that was spun-out of Data61’s predecessor NICTA in 2012, and acquired by IAG in 2015.
The focus of the institute would be to create a “world where all systems behave ethically”. It would do this not just through research, but also through consulting practice and public policy advocacy work.
The institute is being co-led by Bill Simpson-Young, who shifted from Data61 into the new role, and Tiberio Caetano, the current Ambiata chief technology officer. Dr Caetano is also an adjunct professor at Sydney University.
“We are trying to build the science of ethics,” Mr Simpson-Young told InnovationAus.com. “Basically what's happening is that there is a lot of machine learning being used in automated decision systems now, and that has a consequence on people’s lives.”
“The problem that we have is that when people are specifying a problem that they want the machine learning system to solve, it is not specified as well as it should be,” Mr Simpson-Young said.
“And so these people are getting what they asked for: And what they asked for can sometimes have unintended ethical consequences.”
“What we are trying to do is build tools to expose a lot of those trade-offs that are being made in terms of the fairness and the operation of the algorithm, so that people who are responsible for those decision systems can see the implications of what they're doing,” he said.
“[Gradient] is a very technical institute, it’s not a high-level consulting institute ... it’s very much doing new science, building new tools and getting those out into the world to improve the ethical ramifications of automated decision systems.”
The Institute would use research findings to create open source ethical AI tools that can be adopted and adapted for business and government.
It would also conduct fee-for-service consulting services, training programs and public policy advocacy work.
The Gradient Institute partners would not say how much each had invested into the new research institute.