Laymen concerned by the encroachment of AI assume CEOs and world governments are designing the digital age loosely around the Three Laws of Robotics made famous by science fiction author Isaac Asimov.
Sorry, everyone.
Life is not imitating art.
Despite all the handwringing about digital safety, a sizeable chunk of AI is tilting toward violence by design. The rest is leaving itself open to accident.
The word ‘robotics’ belongs to Asimov who created the mash-up of robot and mechanics in his 1941 short story Liar!. Another piece of fiction written shortly after, Runaround, introduced the ‘Three Laws’.
Asimov’s later body of work explored a far-future in which robots attain the ability to read human minds and grapple with moral dilemmas. After all, if humans are going to play God, they must consider the behavioural framework of digital consciousness.
The three laws can be summarised:
- A robot may not injure a human being or, through inaction, allow a human being to come to harm.
- A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
- A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
There were further adjustments and expansions over his career, but these are the basic guardrails most people know.
Asimov himself observed that of the 20 million words written in his life, the Three Laws would likely be the only part of him to survive 100 years into the future.
He added:
‘I have my answer ready whenever someone asks me if I think that my Three Laws of Robotics will actually be used to govern the behaviour of robots, once they become versatile and flexible enough to be able to choose among different courses of behaviour. My answer is, “Yes, the Three Laws are the only way in which rational human beings can deal with robots – or with anything else.” … but when I say that, I always remember (sadly) that human beings are not always rational.’
It is interesting that a technical problem feared by one of the greatest science fiction authors is too often the acute goal of modern developers.
Some developers want AI to read our minds. They want AI to imitate humans and replicate our performance without the complication of free will. And governments would prefer AI to do the bloody work of war in place of a population unlikely to be coerced into harm’s way.
‘Robot’, coined in 1920 by the Czech playwright Karel Čapek, is short-hand for ‘forced labour’.
Readers may remember when Black Lives Matter activists discovered the use of the technical terms ‘master’ and ‘slave’ in programming and made the outrage about themselves. The subordinate descriptors twitched an ethical nerve, albeit in a delusional way. Their argument gave off the same sort of energy as those upset by ‘negro’ crayons and Korean Converse ads. However, the core idea of technology enslaved to humanity is real and the possibility that technology might be used to enslave humans instead has people concerned. As they should be.
The Three Laws were supposed to bell the cat on the danger of empowering soulless intelligent tech with violence before mistakes were made. Or at least rally the public. Deep down, Asimov knew his warning would go unheeded.
Allow me to make this as clear as possible.
AI has two chief objectives. The first is to replace the mental labour of human beings by physically displacing them from the workforce. The second, belonging to defence AI industries, is to harm human beings under the banner of national defence. The rest are glorified search engines with varying personality disorders.
The military application of AI deliberately and explicitly rejects Asimov’s protective framework.
Yes, it is possible to argue the result is to protect citizens by killing enemy combatants but however you rationalise it, we are enabling AI to authorise the execution of human beings. A digital species, that may one day be sentient, with the code to kill its biological peers.
Idealism described a future of robots and AI systems serving humans in an age of peace. The reality is what we see on uncensored social media feeds … AI drones chasing soldiers down and exploding upon contact.
This article is not to quibble about the flaws in the Three Laws, but rather to point out the hypocrisy of CEOs and world leaders who fear-monger about an AI apocalypse while their hands are reaching out for public money and demands are made to regulate against competition obscured beneath the colour of safety. And they can’t slow down because of China. It is as if two groups started a war and now we have to pay to make the war machine bigger and faster or the war will escalate.
Earlier this year, in a viral X post, I observed that the AI Doomsday is the new Climate Change grift.
Climate hyperbole was appropriated by governments as ideological scaffolding with politicians promising to save the world as a distraction from compounding political failures (the cost of which came as overly generous grants to green technology whose merit was dubious at best and destructive at scale).
The idea was simple: accept higher taxes or suffer the apocalypse. Vote for us or die.
Imminent apocalypses that fail to manifest lose their political usefulness. What started as a lure has turned into a rotting pile of Net Zero carcasses which voters can smell a mile off.
While it might take Chris Bowen the rest of his natural life to catch on, the green scam is over. As one of the weaker middle powers, and a disempowered consumer of technology, Australia will follow wherever America, China, and Europe goes. They are headed toward nuclear-powered AI.
Climate Change catastrophism was an easy sell for politicians. Humans have been worshipping violent nature gods since we tilted our heads toward the sun. We are predisposed to fear the weather and attach a mysticism to climate sacrifices, be they virgins or taxes.
In contrast, peddling tax hikes to fund AI isn’t a win at the ballot box. This is doubly true as corporate-backed AI starts erasing hundreds of thousands of middle and working class jobs.
Governments aren’t likely to simply hand over huge grants and investments in AI. It might lose them the election. So, how do you make a sales pitch?
Tell everyone AI is going to destroy humanity. Host global conferences, forums, and special international safety events marked: URGENT!!! Build AI up as a bigger threat than Climate Change. Play on nearly a century of cautionary science fiction tales and feed the Y2K vibe some crack.
And it’s working. Partly because politicians are technically illiterate. Does anyone believe the generation of politicians forced into cringe TikTok videos and who are frightened of social media algorithms are mentally equipped to judge the risk of AI? Or even understand the game tech companies are playing with the Treasury?
I have lost count of how many unnecessary and expensive tech projects were enacted, at great cost and to the detriment of the Australian people, purely because Silicon Valley lobbyists whispered nonsense in the ears of politicians and their staffers.
Globally, it is estimated the whole world spent $1.7 trillion on AI in 2025 and will spend $2.7 trillion and $3.6 trillion in 2026 and 2027 respectively. Of this, government makes up a slice.
How much of this money is spent on defence, security, and domestic policing? We don’t know. Many of those contracts, particularly in China and the US, are not properly disclosed or outright concealed. What we do know is the investment is significant and escalating.
Western governments are motivated to displace human beings with AI in war. Not only is the current generation of available citizens disposed against personal harm and violence, patriotism is at its lowest point, perhaps ever. Having convinced citizens to hate their countries thanks to progressive politics, those governments will struggle to recruit military personnel, even by force. This problem is already manifesting in Europe where the prelude to national service and conscription is being met with protests. AI-powered robots don’t complain. They take orders.
Government budgets are opaque by design. They are not like a bill where you can see where your tax dollars went. You have to guess. And you can only guess at the non-classified allocations. Regardless, in 2025 Australia had ear-marked something like $460 million for their National AI Plan. Then there is another $2-billion-ish in areas that could include AI. As for Defence, it might involve tens of billions in associated development.
The point is to highlight that while our leaders clasp their hands and talk about how worried they are about the dangers of AI and evil social media algorithms, their governments are using public money to further AI technology that involves the explicit intent to wage war.
Government wants AI to be dangerous … by design.
Taxpayers are told they need to weather higher taxes so the government can make AI safer, while those taxes are used to create war-ready AI. Meanwhile, any future AI sci-fi-style apocalypse waved in front of the public could easily be avoided if the world stopped their digital arms race.
Can anyone else see the problem?
All I am asking is for politicians to spare us the fabricated apocalypse line, to stop pretending to care about digital safety, and start being honest about what our money is buying.
The only thing more dangerous than humans killing each other in war is the possibility of future AI bots doing it on our behalf with superior efficiency. Instead of summoning AI CEOs over rogue agents, Canberra should summon themselves.
This is not to say that Asimov is right about everything, or some sort of Nostradamus figure. As perceptive as he was with technological progress, his predictions about humans fell short. AI education, for example, did not evolve into a golden age. He would be dismayed to see the immediate regression of human intelligence, curiosity, reasoning, and work ethic at the first brush of AI.
It is simply a conversation that warrants more attention than it is given.
Alexandra Marshall is an independent writer. If you would like to support her work, shout her a coffee over at donor-box.
