π It was a pleasure speaking at the ASFA | The Voice of Super conference about our ESG-Responsible AI framework (CSIRO’s Data61 Alphinity Investment Management Jessica Cairns )
https://lnkd.in/gw9FFKJG
During my panel, I focused on three key points:
1οΈβ£ The unique nature of this AI wave β Unlike past general-purpose technologies like electricity or computers, AI is different due to its cognitive intelligence (not just a tool) and its multiple purposes right out of the box with minimal further development at times. Plus, itβs based on learning, not fully designed and built by human developers. We need new technical measurements, metrics, and controls more than ever to confidently assure the fulfilment of high-level governance and regulatory aspirations.
2οΈβ£ The nuanced view of AIβs emissions β While training AI consumes significant energy (e.g. one version of ChatGPTβs training was roughly equivalent to the annual power consumption of 60,000+ Australian families), thereβs a twist: the energy consumed by using AI is already surpassing training, as AI becomes more integrated across sectors. For instance, a single chatbot interaction uses about 10x the energy of a search query. But hereβs the second twist: AI can also save emissions downstream, such as reducing hours of emission-heavy work into minutes (Nature article: https://lnkd.in/gnUdhzBa). However, the final twist is that we often donβt stop after saving timeβwe do more, driving up overall consumption and emissions. So, is it worth it? And if so, how do we manage this? β‘
3οΈβ£ AI regulation and standards β I provided an overview of the international landscape, Australiaβs new voluntary AI safety standard, and the consultation on mandatory guardrails for high-risk AI. I couldnβt say much yesterday, but excitingly, itβs out this morning! π https://lnkd.in/gEPVNmVh