Considered policy needed to ensure GenAI works for all workers
RMIT University expert Dr Emmanuelle Walkowiak was among the academics who appeared at yesterday’s public hearing for the Inquiry into the Digital Transformation of Workplaces.
The following is an abridged version of her submission.
Generative AI has properties that are very different from other technologies and will transform our labour market in new ways.
The time to act to ensure that GenAI works for workers is now. My previous research has demonstrated that once technological and organisational choices are implemented, firms do not reverse their choices. Once GenAI is adopted and AI risk mitigation is decided, these choices will not be reversed.
There are two important aspects that I want to flag which differentiate GenAI from other technologies.
Firstly, during previous waves of technology adoption, we have observed a polarisation of income and opportunities on the labour market. What is different this time is that GenAI can enhance the productivity of less skilled and less experienced workers. My research shows GenAI can enhance productivity of less skilled and less experienced workers – positively impacting all workers. Immediate productivity gains for less experienced or less skilled workers was not something observed for other technologies. This is a unique opportunity presented by AI; appropriate policies need to be designed to unlock these opportunities.
Secondly, my recent research into AI risks shows that the transformation of jobs is driven by the inseparability between productivity gains and new AI risks. I analysed the exposure of the Australian Labour Market to eight AI risks (such as privacy or cybersecurity). These AI risks are new types of occupational risks. Mapping of AI risks exposure can help prioritise policy intervention to mitigate these risks.
What does it mean for the labour market? GenAI is a new type of economic agent, involving new types of productivity gains, new forms of learning and new types of risks. The major impact of technological change is a creative destruction of jobs and skills. In the US, research shows that 60% of job titles that existed in 2018 did not exist in the 40s. The creation of new jobs reflects not just automation, but the enabling of new capabilities and services that were not previously feasible without technologies. Policies must be designed and implemented to shape the adoption of GenAI that works for all workers and favour the creations of these new jobs. Upskilling is central to prepare the workforce for the next generation of jobs.
In conclusion, GenAI is about to bring transformational changes in the workplace. These changes can be shaped by adequate policy-making to ensure adoption of AI works for all workers, and there is an opportunity to make AI work for less skilled workers. To ensure that we maximise the benefits of AI adoption, we need a better understanding of emerging risks for workers, which need to be managed with adequate regulations and policies in order to protect workers.
Emmanuelle Walkowiak is Vice-Chancellor’s Senior Research Fellow at RMIT in the School of Economics, Finance and Marketing. She leads the FLOW-GenAI initiative (Future of Labour, Organisation and Work with GenAI).
Like what we do at The AIMN?
You’ll like it even more knowing that your donation will help us to keep up the good fight.
Chuck in a few bucks and see just how far it goes!
Your contribution to help with the running costs of this site will be gratefully accepted.
You can donate through PayPal or credit card via the button below, or donate via bank transfer: BSB: 062500; A/c no: 10495969
1 comment
Login here Register here-
Clakka
Return to home pagePrediction is a beast at the best of times … “garbage in = garbage out”, and “expect the unexpected”. Dr James Lovelock (of the Gaia hpothesis) noted that humans had a flaw, in that they naturally predicted in a time-lineal fashion, when in fact all events over time occurred in an exponential fashion. He noted numerous significant historical examples.
In the mid-1990s I had discussions with a relational economist. He and a network of other relational economists engaged by the UN had over a number of years been gathering and analyzing data from across the world on the march of technology, and its effects on work and society. He advised that their latest conclusion was that within 30-40 years there would only be jobs for about 20% of the population, so the biggest problem for governments would be to keep the remaining 80% amused. It seems we are still waiting.
Understandably, for years there has been concern about the environment and the particularly the climate. But those of the controlling advanced states, typically obsessed about money, have dragged their feet about addressing those concerns, and now it’s reached a level of blind panic. To wit, we have nominated our era as the Anthropocene.
Dr Lovelock in via his Gaia hypothesis proposed that the Earth and its life is a self-sustaining system, where “… oceans and the atmosphere are thoroughly entwined with the living biosphere, and must be understood as a coupled system …”. “Life evolves in response to environmental change, but the environment also evolves in response to biological change.” As time went on, and the human population continued to skyrocket, and the human ravages on the environment increased, Lovelock, in subsequent publications progressively observed the retreat of Gaia’s self-sustainability, and finally settled on an era pivot to the ‘Novacene’, where the Earth will be saved by AI. And it’s notable that in his final years Stephen Hawking effectively said the same.
Hope springs eternal, and as humans generally tend to drag their feet, except in pursuit of money, there’s been a competitive rush by the tech giants to commercialize their AI, via acquisitive ‘chat’ type LLMs (Large Language Models) aka ‘large deep learning models’ – perhaps they should be branded ‘Infant AI’, as they remain a long way from comprehensive AI.
We must remember ‘garbage in = garbage out’, and there’s plenty of garbage about.
Yesterday (3Sept), a Crikey article about an AI test carried out by ASIC, revealed humans outperformed AI on every measure including summarizing and parsing.
There’s much mooted about ‘tipping points’ these days, but is it premature to say we have tipped into the Novacene?