This article is from the source 'guardian' and was first published or seen on . The next check for changes will be
You can find the current article at its original source at https://www.theguardian.com/commentisfree/2025/aug/13/can-ai-deliver-economic-nirvana-only-if-workers-can-monitor-and-shape-how-its-used
The article has changed 2 times. There is an RSS feed of changes available.
Previous version
1
Next version
| Version 0 | Version 1 |
|---|---|
| Can AI deliver economic nirvana? Only if workers can monitor and shape how it’s used | Can AI deliver economic nirvana? Only if workers can monitor and shape how it’s used |
| (about 4 hours later) | |
| It is only when technology augments human capability that genuine productivity is achieved | It is only when technology augments human capability that genuine productivity is achieved |
| One Big Idea is a new series on how to transform Australia’s economy ahead of Jim Chalmers’ economic roundtable | One Big Idea is a new series on how to transform Australia’s economy ahead of Jim Chalmers’ economic roundtable |
| Get our breaking news email, free app or daily news podcast | Get our breaking news email, free app or daily news podcast |
| What is your one big idea? | What is your one big idea? |
| One person I wish was on the treasurer’s roundtable guest list is the reigning Nobel economic prize-winner Daron Acemoglu, whose work lays out a compelling roadmap for the adoption of so-called “artificial intelligence”. | |
| The rapid rollout of these predictive search bots, agents, companions and synthesised creators is being evangelised by the tech industry and our own Productivity Commission as a fast track to prosperity (replace “AI” with the word “God” and the latest PC report reads like a call to prayer). | The rapid rollout of these predictive search bots, agents, companions and synthesised creators is being evangelised by the tech industry and our own Productivity Commission as a fast track to prosperity (replace “AI” with the word “God” and the latest PC report reads like a call to prayer). |
| But Acemoglu’s work shows that when a new technology simply replaces human labour it may well deliver higher profits but few broader economic benefits. It is only when technology augments human capability that genuine productivity is achieved. | But Acemoglu’s work shows that when a new technology simply replaces human labour it may well deliver higher profits but few broader economic benefits. It is only when technology augments human capability that genuine productivity is achieved. |
| To reap that collective dividend, he argues workers must be given the opportunity to not just use the new technology but to shape it: designing new tools, new connections, new markets and new capabilities. | To reap that collective dividend, he argues workers must be given the opportunity to not just use the new technology but to shape it: designing new tools, new connections, new markets and new capabilities. |
| UTS’s Human Technology Institute has tested Acemoglu’s theory with stark results: nurses embrace AI to manage their crippling paperwork but draw the line at patient intervention; retail workers welcome smarter inventory while seeking to preserve customer relationships; gun-shy public servants seek the confidence the technology won’t be used against citizens in another robodebt. | |
| The Australian Council of Trade Unions is calling for stronger protections on AI work replacement, including veto rights. I think Acemoglu would say this is not ambitious enough: if we want AI to add to national wealth rather than extract wealth for tech rent-seekers and lazy cost-cutters, we need to assert democratic control over the technology. | |
| So the big reform idea I’d put forward in Acemoglu’s absence is the mandating of “worker councils” to oversee, monitor and shape the introduction of machine learning and large language models. | So the big reform idea I’d put forward in Acemoglu’s absence is the mandating of “worker councils” to oversee, monitor and shape the introduction of machine learning and large language models. |
| How would it work? | How would it work? |
| Regardless of whether the government ends up pursuing a standalone AI act, it should expand the general duty of care employers discharge under workplace safety to include the introduction of new technology. | Regardless of whether the government ends up pursuing a standalone AI act, it should expand the general duty of care employers discharge under workplace safety to include the introduction of new technology. |
| To discharge that general duty, employers (either individually or as an industry) would need to show they had actively engaged their workers in the technology and had given them the opportunity to test, refine, propose guardrails and enforce red lines. | To discharge that general duty, employers (either individually or as an industry) would need to show they had actively engaged their workers in the technology and had given them the opportunity to test, refine, propose guardrails and enforce red lines. |
| The workers councils would be democratic and representative, provided with the information needed to understand the technology, the remit to observe its use in the workplace and an ongoing role in monitoring its impact. | The workers councils would be democratic and representative, provided with the information needed to understand the technology, the remit to observe its use in the workplace and an ongoing role in monitoring its impact. |
| Where a union is active this could be through existing consultative processes; where there is no union, it would be incumbent upon employers or broader industry bodies to establish genuinely accountable processes. | Where a union is active this could be through existing consultative processes; where there is no union, it would be incumbent upon employers or broader industry bodies to establish genuinely accountable processes. |
| This needs to be part of the broader discussion about the way AI is regulated, rather than looking for specific rules for future use cases establishing that a general duty of care creates ongoing accountability for when things go wrong (as they inevitably will). | |
| What are the downsides? | What are the downsides? |
| Employers would cry foul about more “red tape” or a worker takeover of business operations; but actively involving workers is not about giving up power; it is harnessing existing knowledge. | Employers would cry foul about more “red tape” or a worker takeover of business operations; but actively involving workers is not about giving up power; it is harnessing existing knowledge. |
| Let’s be honest: most tech transformations are complete failures. Not because the tech itself is bad, but because it’s just not suited to the realities of how people work. To think that AI is somehow different in this regard is what the vendors want, but it’s madness. | Let’s be honest: most tech transformations are complete failures. Not because the tech itself is bad, but because it’s just not suited to the realities of how people work. To think that AI is somehow different in this regard is what the vendors want, but it’s madness. |
| A broader pushback will come from the tech industry, which is pitching the narrative that we need to “win” the race to AI before other countries get ahead of us, meaning any move to impose friction around change risks “stifling innovation”. | A broader pushback will come from the tech industry, which is pitching the narrative that we need to “win” the race to AI before other countries get ahead of us, meaning any move to impose friction around change risks “stifling innovation”. |
| If democracy wasn’t facing collapse, if our kids weren’t falling prey to exploitative algorithms, if the models weren’t built on the stolen labour of creators, if the same companies pushing AI hadn’t abrogated their responsibility with social media platforms, maybe we would give them the benefit of the doubt. But they have not earned this. | If democracy wasn’t facing collapse, if our kids weren’t falling prey to exploitative algorithms, if the models weren’t built on the stolen labour of creators, if the same companies pushing AI hadn’t abrogated their responsibility with social media platforms, maybe we would give them the benefit of the doubt. But they have not earned this. |
| There may be advantages in stepping back and not racing headlong into a future driven by self-interested and increasingly bad faith actors; refusing to allow our workforce to be their canary in the coalmine, with government expected to step in when things go pear-shaped. | There may be advantages in stepping back and not racing headlong into a future driven by self-interested and increasingly bad faith actors; refusing to allow our workforce to be their canary in the coalmine, with government expected to step in when things go pear-shaped. |
| Rather than be the invisible bystanders of Big Tech, workers, through AI councils, can become part of a hard-headed economic response to a technology that is still being developed, with models and structures that are still in a state of flux. | |
| How about the politics? | How about the politics? |
| We know Australians are deeply sceptical of AI. There is a growing sense that it is outside of our control and being imposed on us for some ill-defined greater good. | We know Australians are deeply sceptical of AI. There is a growing sense that it is outside of our control and being imposed on us for some ill-defined greater good. |
| Imagining new sorts of democratic structures would give us a modicum of agency in determining how these tools should evolve and how the core resource driving them – our data – should be collected, refined and how we should be compensated for its use. | Imagining new sorts of democratic structures would give us a modicum of agency in determining how these tools should evolve and how the core resource driving them – our data – should be collected, refined and how we should be compensated for its use. |
| Indeed, the same structures that would govern AI in the workplace could be developed to give citizens a broader voice in the development of technology that is still highly contested. | Indeed, the same structures that would govern AI in the workplace could be developed to give citizens a broader voice in the development of technology that is still highly contested. |
| Our tech future is not predetermined; ensuring we all have an informed say in these hugely consequential decisions will not just define the technology, but the sort of nation we become. | |
| And if the apostles of AI are right and this really is godlike technology that will deliver economic nirvana, then embedding these democratic structures in their development will help earn the trust of a public that has been over-sold on tech utopia before. | |
| Peter Lewis is the executive director of Essential, a progressive strategic communications and research company that undertook research for Labor in the last election and conducts qualitative research for Guardian Australia. He is also the host of Per Capita’s Burning Platforms podcast | Peter Lewis is the executive director of Essential, a progressive strategic communications and research company that undertook research for Labor in the last election and conducts qualitative research for Guardian Australia. He is also the host of Per Capita’s Burning Platforms podcast |
Previous version
1
Next version