analysis
There are big opportunities in AI, but also very real risks — how can governments get the balance right?
Artificial intelligence has already moved from sci-fi fantasy to everyday use. (Reuters: Dado Ruvic/Illustration)
When Malcolm Turnbull talked up his "innovation agenda" after seizing the prime ministership in 2015, instead of sparking excitement about the future, all that many voters heard was that their jobs were on the line.
Innovation became code for dangerous change, according to Liberals on the frontline. While talking about tech and radical transformative change might sound bold and forward-thinking, a nervous electorate with change fatigue often only fears that life is about to get harder. Unless you take them with you on the change journey, cynicism prevails.
The truth is politics is consumed with the every day — from tax cuts, to housing prices and cost-of-living relief. But a quiet revolution is going on that threatens to change everything in our lives and oddly there is disturbingly little national debate about what we should do about it.
Our political leaders haven't really spent much time talking about it — but they should be and they should be engaging us.
Social media got away
On Q+A last Monday, former Australian Financial Review columnist and now author Joe Aston compared social media giants to big tobacco. The reality is that social media blew up so fast and accelerated so hard the consequences were barely debated.
We gave away our privacy, we overshared and allowed algorithms to tell us what we wanted and what we liked without really reckoning with what it meant and how we wanted to use it. We only started having deeper ethical deliberations after we were already addicts and had been mugged by the pace of change.
Now we are immersed in a debate about whether we should follow the US in banning TikTok because of its data vulnerability in China, but the horse has, in many ways, already bolted. Millions of young people are not only obsessed with the platform, but many have monetised it, making it a brave government that stands in the way of that basic economic reality.
In Australia, we risk doing more of the same unless we turn our collective attention to the benefits and risks of artificial intelligence in a meaningful way.
The new frontier
Artificial intelligence has already moved from sci-fi fantasy to everyday use — from facial recognition on our phones to the $100 billion ChatGPT program which generates human-like responses to every possible instruction.
I considered asking AI to write this column just to make the point. On Q+A most weeks, we've got audience members telling us they are using it to ask their questions. The opportunities are infinite, as are the concerns, from deepfakes with the potential to swing elections and damage democracy to workers being made redundant, which is an uncontested inevitability. All of it is happening in real-time. We just haven't worked out what we will accept and what we will repudiate.
When a group of industry leaders warned last year that the AI technology they were building might one day pose an existential threat to humanity and should be considered a societal risk similar to pandemics and nuclear wars, the world finally started to take notice.
"Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks, such as pandemics and nuclear war," reads a one-sentence statement released by the Center for AI Safety, a nonprofit organisation. The open letter was signed by more than 350 executives, researchers and engineers working in AI.
That statement changed the game and ever since the world and nation-states like ours have been engaged. But is the pace of that engagement enough, and how do we get the balance right?
Moves to regulate AI
Last week, the European Union voted on world-leading legislation that says it will protect citizens from the possible risks of AI technology developing at incomprehensible speeds, while also maintaining innovation across Europe. Yep, there's that word again.
The EU wanted to pass the new laws since ChatGPT came along in late 2022, launching an AI race across the planet. It changed the game so quickly that most of our heads are still spinning.
The Albanese government has announced the creation of a new AI expert group to offer advice on what laws here will look like.
The group first met in February and will continue offering advice until June 30 this year. The government has said they will introduce a risk-based system to protect against the worst potential harms of AI and that risky technologies will have mandatory rules applied to them, including even independent assessments and audits. The government has said they will avoid stifling the growth of low-risk AI.
Industry and Science Minister Ed Husic, who is deeply exercised about this issue, has flagged the parameters of new laws, including plans for AI-generated content to be labelled so it can't be mistaken as real.
But even before the new laws have been finalised, AI is disrupting life as we know it.
The Sydney Morning Herald revealed last week that universities are catching hundreds of students in alleged cheating using ChatGPT or other artificial intelligence, as universities use new anti-plagiarism tools to catch people out.
An International Monetary Fund study recently predicted that AI was poised to impact about 60 per cent of all jobs in advanced economies like ours, with about half of those to benefit from AI increasing productivity and the other half to be adversely impacted. That means job losses.
ChatGPT came along in late 2022, launching an AI race across the planet. (ABC Gold Coast: Dominic Cansdale)
Husic wants the advice of his expert group on what he has called "mandatory guardrails" for high-risk AI settings. "With expertise in law, ethics and technology, I'm confident this group will get the balance right."
Husic says that at its best, AI can help improve the quality of life and how our economy functions. "At its worst, it can distort and divide. I have concerns about how biometrics are used, social scoring for example," he says.
"And of course, the use of AI which seeks to manipulate actions with a criminal intent."
Husic says the government has already got started on key initiatives.
"We've established a short-term expert advisory group to help guide the development of mandatory guardrails for high-risk AI, they will report back in June.
"Mandatory regulation for high-risk AI could be in place as early as the end of the year."
Husic says his strategy is to build trust and transparency in AI systems and he is working with industry to develop a voluntary AI safety standard, including labelling and watermarking of AI-generated material.
But he acknowledges the threat to jobs. "As with the introduction of any new technology, we need to be aware of the potential impact on employment, keeping in mind AI will help to build new industries and create well-paid secure jobs.
"We want to allow low-risk AI to flourish unimpeded."
LoadingRegulation should be clear and focused
Last November, Husic attended the UK AI Safety Summit where Australia — along with 27 other countries — signed up to the Bletchley Declaration. He says it was a big step toward creating a safe and trustworthy international framework to tackle the risks of AI.
"Australia is closely monitoring how other countries are responding to the challenges of AI, while getting cracking on its own regulatory framework. The days of self-regulation for technology are over," he says.
Loading...The opposition says they are aware of the risks of change but are also troubled by heavy regulation and reluctant to overly regulate in a way that stifles businesses that want to adopt AI to make tasks faster and easier. They are now starting to voice their parameters for what the mid-year legislation should look like.
Shadow communications minister David Coleman told ABC News the opportunity in AI is "bigger than in any technology since the creation of the internet itself". He says it's crucial that we develop and own AI intellectual property here in Australia. "The government needs to help make that happen by encouraging investment. Much more action is needed on this."
On regulation, Coleman says we need to be very clear and focused on where we should act and where we should not. "Most aspects of AI do not require regulation, and excessive government involvement would make Australia less competitive. While there are some positive aspects of the EU AI law, it errs too much on the side of regulation," he says.
"But there are very real risks from AI that will require regulation. We must ensure that sovereign nations always determine the red lines that must not be crossed by software. Australia should be playing a leading role in working with international partners to determine these rules. At present, there is little sign that the Australian government is providing intellectual leadership on these issues."
Before the government does anything to regulate and legislate, we need to have our own reckoning as a community about how we want to use this technology in our lives or face having those decisions made for us again.
Patricia Karvelas is the presenter of Q+A, which airs at 9.35pm tonight on ABC TV and ABC iview, RN Breakfast and co-host of the Party Room podcast.