What I’d do if AI came for my career at 45

I keep coming back to the arithmetic of 45. It is old enough to have done some things right and young enough that you cannot simply opt out. A salary of $120,000 sounds solid. Savings of $300,000 sound even better. Then artificial intelligence arrives as a workplace rumour, a board deck, a new software tab, and suddenly that tidy little story about being “set” stops working.

Patrick’s case on Yahoo Finance / Moneywise has stuck with me for that reason. He is 45, earns $120,000, has $300,000 saved, and still feels as if his career could be hollowed out within a few years. The numbers matter, of course. So does the psychological shift. Once you suspect the field itself may be shrinking, savings stop reading as freedom and start reading as runway. The same piece points to estimates that 41% of employers expect to shrink their workforce because of AI automation and 60% of workers may need retraining by 2030. Even if those figures never land neatly in your office, they change how the week feels.

Still, the corporate line in Australia is calmer than that. Australian CEOs favour AI reskilling over job cuts, according to reporting from IT Brief, with EY-Parthenon’s Shannon Cotter arguing that the serious money is now going into redesigning jobs rather than simply deleting them. I do not think that view is fake. I do think it can sound oddly serene if you are the one sitting in a swivel chair wondering whether your useful skills are about to become background noise.

When the savings stop feeling like safety

For mid-career workers, AI anxiety is rarely just fear of technology. It is fear of timing. At 25, you can blow up a plan and call it reinvention. At 45, you are more likely to hear a very practical question thudding in the background: how long would I have if the role disappeared, and who exactly would hire me next?

Local reporting on white-collar redundancies matters more to me than the wilder Silicon Valley rhetoric. In news.com.au’s reporting on lawyers fielding AI-linked job loss cases and in the Guardian’s reporting on WiseTech redundancies, the mood is procedural and immediate. Emails go out. Roles get folded into “transformation”. Workers are told the business is evolving. Then somebody still has to pay rent in June.

“This is what AI disruption looks like on the ground and workers are terrified.”
— Paul Inglis, news.com.au

From the labour side, Inglis describes the bodily version of the problem. You can believe the long-run productivity story and still admit that redundancy feels awful in the month it happens. Those two truths can sit together. In fact they probably have to.

Oddly, the perspective that steadies me is the skeptical one. MIT Technology Review argued this week that there is still scant evidence AI has caused a large-scale labour-market shock in the United States. Business Insider also reported on research suggesting work from home has been a stronger driver of weaker junior hiring than generative AI exposure. That does not erase the anxiety. It does answer one useful question from the fact bundle: are we already looking at clean evidence that AI alone is smashing the jobs market? No. The evidence is messier than that. Companies are cutting in some places, hiring oddly in others, and using AI as one explanation among several. If you are 45, that mixed signal is not exactly comforting, but it does suggest panic is a poor strategist.

To me, a slower unbundling of white-collar work feels more believable. The repetitive parts go first. Drafting, admin, first-pass analysis, the parts of a role that once made a junior worker worth hiring or a mid-level worker seem hard to replace. Judgment survives longer. Context survives longer. So does trust. The trouble is that these are harder to itemise on a CV and harder to defend in a quarterly budget review. A solid balance sheet can still feel flimsy when the thing under pressure is not your employer this quarter, but your function over the next five years.

The plan I’d trust more than panic

For Patrick, or anyone else doing the eerie maths of mid-career risk, I would start by downgrading the fantasy that one big decision will solve this. It will not. There is no single course, no magical AI certificate, no neat moment when the market taps you on the shoulder and says congratulations, you are future-proof now. The useful response looks embarrassingly ordinary. It starts with cash, then skills, then relationships, in that order.

Cash comes first. I would work out exactly how many months $300,000 really buys once tax, housing costs and simple bad luck enter the frame. I would want the number. A vibe is useless at 3am. Mid-career workers often talk about savings as if the existence of the pot is the comfort. In practice, the comfort comes from knowing the burn rate. Runway is a planning tool. Vagueness is what keeps you awake.

Skills are next, and here I think the employer view is worth hearing in full. In IT Brief’s reporting on Australian CEOs leaning toward reskilling, Cotter put it this way:

“CEOs are investing more, and they want a return. That means redesigning work, supporting people through change and measuring outcomes, not just running pilots.”
— Shannon Cotter, IT Brief Australia

That is a much better sentence than the usual fluff about innovation. It implies companies are done treating AI as a toy. They want workflows to change. If I were trying to stay employed through that shift, I would stop asking whether AI can replace me and start asking which 20% of my week it can already do faster, and which 20% of my week becomes more valuable once that grunt work disappears. Somewhere in that answer is the next version of the role.

I would also treat AI literacy as a routine career expense, the same way previous generations learned Excel properly or got comfortable presenting to a room. Business Insider’s practical roundup on AI-proofing a job is not gospel, but its underlying point is sound: audit the tasks before the title. Job titles survive longer in conversation than they do in budgets. Task bundles are what get reorganised.

Emotion is the part where a lot of otherwise sensible career advice falls apart. People hear “retrain” and imagine a cheerful adult-learning montage. Real life is uglier. Retraining takes time when your time is already sold. It takes humility when your identity is tied to being competent. It often takes money before it saves money. That is why I keep coming back to The Conversation’s piece on job insecurity, which treats the feeling itself as part of the story rather than a private weakness to be managed off to the side.

“But there are evidence-based things we can do – at an individual, organisational and government level – to manage job insecurity better in uncertain times.”
— Sharon Kaye Parker, The Conversation

That, to me, is the adult framing. Mid-career workers do not need another sermon about resilience. They need employers that train before cutting, governments that make genuine re-skilling less punitive, and enough personal clarity to admit when a field is changing faster than loyalty can fix. InnovationAus recently captured the policy version of that tension: Australia wants the upside of AI without a jobs panic. Fair enough. Workers want the same bargain.

Finally, I would get less romantic about professional identity. If a field is becoming thinner, I would widen the circle around what counts as “my kind of work”. That may mean adjacent roles, contract work, a sideways move inside the same industry, or a year spent building a newer skill while the old one still pays the bills. None of that sounds glamorous. That is partly why it works. Career preservation in your forties is usually not a dramatic pivot. It is a series of mildly unsexy decisions made before the cliff edge arrives.

Nor do I buy the loudest version of the AI story, the one that says your career is either untouched or doomed. Most working lives are not that clean. What I hear in Patrick’s numbers is something more familiar, and more useful: a person realising that decent savings do not cancel uncertainty, they buy time to respond to it. Time is not safety. Still, it is something. Used well, it may be the difference between being replaced by a story about the future and quietly organising your place inside it.