I Wrote This in 2014. I Called Them "Invisible Robots."
You probably have a different name for them now.
A decade ago I wrote a piece warning that a new generation of automated systems was about to come romping into the workplace — invisible, efficient and profoundly disruptive — Robotic Process Automation. I said organisations were hugely unprepared. I said the human impacts would be systematically underestimated. I said leaders who avoided honest conversations with their people about displacement would store up far larger problems down the track.
I underestimated the scale and the speed by an order of magnitude.
What I described as "invisible robots" in 2014 have become a plethora of large language models. The transitional technology I predicted has arrived — followed by so much more. ChatGPT reached 100 million users in two months, faster than any technology in history. Microsoft deployed AI co-pilots across the most widely used productivity suite on earth. Goldman Sachs estimated in 2023 that approximately two-thirds of current occupations are exposed to some degree of AI automation, with 300 million jobs globally at meaningful risk of disruption. The wave I saw coming from a distance is now breaking over the shore.
So this is not a new story. It is an updated one. The warnings are the same. The urgency is considerably higher.
The optimistic bias has not gone away
The most acute observation I made in 2014 came from Pew Research, which found that while two-thirds of Americans believed robots would inevitably perform most human work within fifty years, about 80% also believed their own jobs would survive unchanged over the same period. That finding has been replicated in study after study since, including specifically around AI. People will readily acknowledge that artificial intelligence will transform their industry while simultaneously believing — with apparently no sense of contradiction — that their own role, their own team and their own organisation are somehow exempt.
This is not stupidity. It is one of the most reliable features of human cognition when confronted with large-scale, abstract threats. It is also one of the most dangerous postures a leader can adopt in 2026 and beyond.
The scale is not abstract anymore
The World Economic Forum's most recent Future of Jobs report projects 83 million jobs displaced by automation over the next five years, against 69 million newly created — a net loss of 14 million roles, concentrated heavily in administrative, clerical and operational functions. These are not factory workers. They are knowledge workers: the analysts, coordinators, customer service specialists and process managers who make up the middle layers of most large organisations.
The jobs that AI is targeting first are not the boring ones nobody wants. They are the entry-level professional roles that used to represent the bottom rungs of a career ladder — the positions where people learn an industry, build networks and develop judgment. Automating them delivers short-term efficiency gains and long-term capability deficits that most organisations have not begun to account for.
We still need Asimov's compact
In 1942, Isaac Asimov introduced his Three Laws of Robotics in a piece of science fiction. A robot may not injure a human being. A robot must obey human orders. A robot must protect its own existence. They were written as fiction, but they described a real governance challenge that the AI industry is still failing to resolve adequately in 2025.
What I argued in 2014 — and what I believe more strongly now — is that organisations need their own version of that compact with employees. Not a legal document or a policy framework, but a genuine, ongoing conversation about what AI means for the people doing the work: which tasks will be automated, what new roles will emerge, how the organisation intends to invest in the transition, and what the honest timeline looks like.
Most organisations are not having that conversation. They are running AI pilots, publishing AI strategies and forming AI steering committees while the people most affected are reading between the lines of internal communications and drawing their own, usually pessimistic, conclusions.
The candid conversation is not optional
Strong leaders — the ones I have consistently seen navigate technology transitions successfully — do not wait until the implications are undeniable before engaging their people. They name the disruption early, acknowledge the uncertainty honestly and invest in creating genuine pathways for those who will be displaced.
This is not sentiment. It is strategy. An organisation whose people trust that leadership will deal with them honestly through a difficult transition performs the transition faster, retains more institutional knowledge and emerges with a healthier culture than one that manages the same change through reassurance and opacity.
I said this in 2014. I am saying it again now, for the same reasons and with considerably more evidence behind me.
The robots have arrived. They are not called robots anymore. And we are still, as an enterprise leadership community, largely unprepared.
