Save articles for later
Add articles to your saved list and come back to them any time.
Before a new drug is released to the market it undergoes significant trials and tests. It’s subject to regulation, assessed by a government agency, prescribed by doctors and dispensed at chemists.
Closer to home, before your next-door neighbours can add a second storey to their house, it goes through a planning approval process to determine its structural efficacy, and to interrogate the impact it will have on the amenity of the neighbourhood and adjacent properties.
The power of AI is unknown, which is why we need regulation.Credit: Marija Ercegovac
As a community, we expect checks and balances – for safety, wellbeing, and risk mitigation. Curiously, most of us blithely embrace new technology without the same concerns or, it seems, with any expectations of regulatory oversight.
It is time to question why, on one hand, we have come to trust unregulated tech, but on the other hand acknowledge that it would take the most extreme circumstances for us to consider buying medicine for a sick child from a shaman purporting to offer a cure stashed in the boot of their car.
Many of the tech products we consume daily are free, readily available, easy to use, and seem harmless. Because of that, we have let Facebook and Instagram observe our private lives, TikTok capture our goofy dance videos with our kids, and Google and Bing take note of everything we search.
Now products such as ChatGPT, Bard, and other generative Artificial Intelligence tools and large language models have entered our lives. As with their older tech cousins, we have allowed a smarter and more powerful stranger waltz into our home without having to show any credentials.
Let’s be clear: these products are not harmless. Although many esteemed academics and researchers have written and spoken about the risks of AI for years, it was only when one of the “godfathers” of AI, Geoffrey Hinton, grabbed the conch shell that the broader community started to listen. Hinton, a neuroscientist and computer scientist, quit his job at Google recently to share his concerns.
Imagine for a moment Hinton worked in a lab that created a flu vaccine, which two years later he admitted to the world caused birth defects. All hell would break loose. Instead, Hinton and others created amorphous tech products, beyond the imagination of most of us, which he now acknowledges pose an existential threat with serious consequences for humanity.
Since leaving Google, Hinton has said that “large language” models such as ChatGPT will lead to a rise in fake news, disinformation, divisiveness, and the feeding of information that will spur righteous indignation like a drug.
Although generative AI is not (yet) posing a direct threat to our physical wellbeing, left to its own devices, it could be as dangerous for society as an unregulated drug or an illegally constructed apartment block.
We have spent years observing the advantages afforded to tech giant first movers. These early pioneers have secured their land and built almost impenetrable moats around their castles. They are deeply embedded in our communities and our psyche. We should be in no doubt that they will continue to use their expertise and financial muscle to lead the AI arms race, and it should not happen unchecked.
Pleasingly, governments – including our own – are starting to sit up and take notice of the risks. The Albanese government’s Safe and Responsible AI in Australia discussion paper presents an opportunity for concerns to be raised and fresh ideas to be ventilated.
We should be considering early intervention – preferably before these products hit the market. We don’t need to understand how or why Panadol reduces our headache because we know it has been through rigorous testing.
The opposite is true for AI; we don’t understand how it works and that’s why we need better testing and appropriate legal protections before it is released on an unsuspecting public. And at very least we should look at imposing corporate penalties, akin to director liabilities that exist under Australia’s work health safety laws so that those at the helm of companies creating these products are forced to pause and reflect.
Of course, there are existing laws that can address some the challenges posed by AI. We should get on the front foot, use those tools now and resource our regulators well.
With big tech’s track record of monopolistic power, accountability problems and transparency failings, there is a clear need for governments around the world to step up to rebalance the power and protect the interests of everyday citizens.
Emma McDonald is director of Minderoo Foundation’s Frontier Technology initiative. Minderoo Foundation supports work on AI research through partnerships with UTS, UCLA and Cambridge University.
The Opinion newsletter is a weekly wrap of views that will challenge, champion and inform your own. Sign up here.
Most Viewed in Technology
From our partners
Source: Read Full Article