As reported by RNZ via ABC News, a 1968 Four Corners television report has resurfaced in which journalist Jim Downes asked Australians: “When this circuit learns your job, what are you going to do?” The AI jobs threat in 2026 is drawing direct comparisons to the fear and disruption that computers generated sixty years ago, with academics, regulators, and business leaders debating whether AI will ultimately destroy jobs or simply transform them. The parallel is striking: the same arguments being made today about AI were made about computers in 1968, and the experts of that era were largely wrong about both the speed and the scale of disruption.
Key Insights
- In 1968, fewer than 600 computers existed in Australia; experts predicted entire professions would be wiped out within five years – neither prediction came true as forecast
- Experts predicted that by 1980, only 10% of the workforce would be needed to maintain economic output
- Australia’s Department of Employment found the job market remains strong despite thousands of tech worker layoffs linked to AI, with early weakening only around roles most exposed to AI such as telemarketers
- South Australian Premier Peter Malinauskas has announced a royal commission into AI, calling it both an opportunity and, unchecked, a material risk to the future of work
- Professor Toby Walsh, UNSW: AI disruption will likely “take longer than we expect”, consistent with every major technology wave before it
- Professor Justin Zobel, University of Melbourne: AI is the first technology capable of “difficult knowledge work” – a genuinely different category of disruption from all previous automation
Our Thoughts
The AI jobs threat 2026 debate is, as UNSW’s Professor Walsh puts it, a conversation that seems to be repeating itself. In 1968, the arrival of the computer triggered the same mix of excitement, anxiety, and wildly overconfident prediction that AI is generating today. Experts predicted that clerk officers had five years left. They predicted that by 1980, a tenth of the workforce would be sufficient to run the entire economy. Neither prediction came close to being right. The technology was real, the disruption was real, but the timeline and the scale were consistently overstated by the people closest to the technology.
This is not a reason for complacency. It is a reason for calibration. The pattern Walsh identifies, that technology tends to take longer to reshape institutions and workplaces than early adopters and advocates predict, does not mean the disruption never arrives. It means businesses and workers have more time to adapt than the most alarming headlines suggest, but only if they use that time deliberately rather than assuming the threat has passed. The internet connected Australia in 1989. Broadband did not become widespread until the late 2000s, nearly twenty years later. The transformation it produced was nonetheless profound and irreversible. AI is likely to follow a similar arc: slower than the hype, faster than the sceptics, and ultimately more consequential than either camp is currently willing to say.
The distinction Professor Zobel draws is the most important one for business owners to hold onto. Previous technologies largely automated physical and routine cognitive tasks. AI is the first technology capable of performing what he calls difficult knowledge work: analysis, writing, legal reasoning, financial modelling, medical diagnosis. That is a genuinely different category of disruption. The jobs most at risk are not the ones we typically associate with automation. They are the ones that have historically been considered safe because they required education, judgement, and expertise. If you run a professional services business, a law firm, an accounting practice, a creative agency, or any other knowledge-intensive enterprise, the AI jobs threat in 2026 is not a manufacturing sector problem that does not apply to you.
And yet the 1968 parallel holds a genuinely reassuring thread alongside the cautionary one. Trevor Pearcey’s prediction from sixty years ago remains the most accurate: some people will become redundant, but there will always be new things they can do, even if retraining is required. The question is not whether AI changes the nature of work. It will. The question is whether individuals and businesses invest in that retraining proactively, before redundancy forces the issue, or reactively, after the disruption has already arrived. The businesses that prospered through the computer revolution were not the ones that resisted computers. They were the ones that figured out how to use computers better than their competitors.
For Black Arrow’s audience, the practical implication is not to panic and not to ignore. The most useful thing a business owner can do right now is identify which tasks in their operation are most exposed to AI capability, which are most dependent on human judgement and relationship, and what the business looks like if the former category becomes substantially cheaper or faster to perform. That exercise is worth doing now, while the pressure is still manageable, rather than in five years when the market has already moved.
Our Questions for You
- In 1968, experts predicted computers would eliminate entire professions within five years. Those predictions were wrong about the timeline, if not ultimately the direction. Do you think the experts predicting AI will displace knowledge workers within a decade are making the same mistake, or is this time genuinely different?
- Professor Zobel argues that AI is the first technology capable of doing difficult knowledge work, which sets it apart from all previous automation waves. If that is true, which specific tasks or roles in your business do you think are most exposed, and what are you doing about it?
- The 1968 parallel suggests that the businesses that thrive through technological disruption are the ones that adopt the new technology faster than their competitors, not the ones that resist it. What is the single biggest barrier stopping your business from moving faster on AI adoption right now?
