AI: The existential threat that’s already here

I recently caught a flight home to Sydney. When I landed, I collected my bags after being alerted they were about to be deposited; knowing they weren’t yet on the carousel gave me time to drop into a shop. I usually look forward to some casual cashier interaction, but there was no one there; instead, I paid at the self-service. Having completed my airport experience, I picked up my phone and ordered an Uber, a pizza and beer to meet me at home.

I am sure the journey I described is something that most, if not all of you, have repeated versions of. It’s normal and forgettable. Having been born during a time when literally all of the above was done by people, I can tell you without a doubt that tomorrow’s world is much better, for convenience, speed and safety. But at what cost?

The Organisation for Economic Co-operation and Development (OECD) reported in 2021 that of 21 countries they investigated, each country saw growth despite how many roles were automated. Those countries that had the highest potential for automation grew fastest due to productivity gains from automating routine tasks.

The report also found that low-skilled and older people did find other work, but often found themselves concentrated in roles that could be replaced by automation. An example of this automation is self-service checkouts, which were introduced in Australia in 2008. With transaction speeds increasing as much as 10% and with up to 65% of customers preferring using them, it’s clear that the right automation can drive revenue and customer satisfaction, even if you need to “place the item in the bagging area.”

McKinsey investigated this in their 2025 Superagency in the Workplace report. They found existing AI technologies could automate activities that absorb up to 70% of employees’ time. This should give you pause…and it’s already at your peripheral. Adobe provides a free summary tool for PDFs. CoPilot allows for a post-holiday summary of missed emails. If this trend holds true to the low-skilled automation findings from the OECD, those businesses that embrace these AI tools will see revenue grow.

PWC agreed in their AI jobs barometer: Industries with exposure to AI saw revenue per employee increase three times faster compared to those least exposed; 27% versus 9%. The good news for these workers is PWC assesses a 56% growth in wages, which was double the growth rate of the previous year for AI-exposed employees.

The scale of workplace change was put clearly by McKinsey, which estimates that up to 14% of the global workforce may need to change professions by 2030. This demonstrates, alongside the OECD report, that those workers who can pivot or train to do a different profession will fare the best. Again, though, drawing on the OECD’s report findings, it’s likely that those who cannot adapt will stay stuck in roles with a high risk of automation. Consider this like a game of musical chairs: With each pause of the music, more roles disappear.

Advice for parents

Rather than encourage your children into the career paths you took, consider the need for AI in every role, not simply as an assistant, but often replacing humans completely. Take Harvey, the law firm started in 2022 by junior lawyer Winston Weinberg. It’s now worth $5B after securing $300M in series E funding. With fewer chairs to sit on, your kids will need more than great grades. Employers will have the brightest humans to choose from. While it’s hard to say which skills will be in demand in 15 years, it’s clear which traits will make people successful: Grit, laser-like focus, a deep specialization and enthusiasm for learning.

Currently, a graduate can learn new skills while pursuing a degree and expect to use them over the course of a career. That era is over; they will need to be lifelong learners.

I suspect there will always be a need for human-centric connection in certain roles. While graphic artists might struggle to keep pace with the demands of the computer game industry, there will always be spaces for the next Andy Warhol or Jackson Pollock.

But we shouldn’t assume that just because a role has a human element that it will be safe. Take teachers as an example; already, schools in developing countries like Ghana and Rwanda are experimenting with AI tutors. Your children may go to the very best human-led schools now, but what if they could get access to the best AI teaching in the world, at a time convenient for you, that achieved better results than your local human school?

Are you preparing your children for a world where the best teachers are human? How long will that paradigm last? Like King Canute, you shouldn’t stand on the shoreline trying to hold back the tide, but instead build an ark.

Advice for business

The path to innovation has always been littered with best intentions, innovation, false starts and failures. AI is no different, but the speed of failure is faster. AI adaptation is moving quickly; according to McKinsey, in 2024, 65% of organizations regularly use AI, up from 33% in 2023, almost doubling in a year.

There are two paths to AI: planning and strategy. Planning starts with defining your service catalog; this will help you spot areas where AI might make sense. From there, translate it into use cases and decide which services you need to remain human. With these in place, you can decide which route gives you the best return: process efficiencies, outsourcing or automation.

Only once you’ve matured the first three should you move into AI. You’ll also need a C-suite leader with a board mandate to own your AI transformation. In most businesses, that’s likely to be the CTO. Then embed AI business partners into at least finance, HR and your product teams. Technical people who know what is possible and can turn that into use cases that reduce headcount, increase revenue and improve EBITDA.

The strategic path is simpler. Companies should carve their business into what makes them unique; their key differentiators or unique service offerings. Anything tangential or best left to partnerships — i.e., not part of the core — should be considered for outsourcing or AI transformation. Short term, most companies that adopt AI in their core will fail to realize the promised gains; focus on what your business is, keep it human and cut away the rest.

For those who go down the hard road of building models, remember the lessons of past tech debt; AI infrastructure is capital-intensive and difficult to maintain. SaaS tools, by contrast, are operational spend and usually the smarter path, but they take directly from your in-year profit. Only you and your finance team can decide what’s best. That said, when compared with how human labor is costed in shared services or projects, the choice is often like-for-like: capex or opex; people or platforms.

Whatever you decide to do, waiting is not an option. If you’re the CEO, don’t wait to work this out or for someone else to bring this to you. Form a task group to find how AI can cut costs, increase value or open new markets.

Advice for governments

In a world where most jobs are automated or replaced with AI, how do we live? How do we retire? Will our judges and lawyers be AI? Will our politicians? These are big questions. Get the answers wrong and the consequences will be far-reaching. Voters deserve manifesto promises about the future of work and the economy. It’s down to the political parties to answer the hard questions, not the mandarins.

If a company chooses to release people in favor of AI or automation, should they pay a year’s salary and retrain them out of the automation trap? For workers close to retirement, should companies pay the difference so they can retire early? If any of these workers falls back into a role that is vulnerable to automation, then could this money be clawed back? Suggestions like these might go some way to address the issues found in the OECD report.

Throughout human history, we have demonstrated an extraordinary ability to remove the agency of fellow humans. This time, we have an opportunity before the first sentient AI arrives to get it right. If an AI is born into a company, who is responsible for its welfare and crimes? Globally, governments must legislate now to set minimum rights for sentience to protect these proto-minds.

Just like my flight home, this technology will soon be part of our everyday lives, and our children will grow up in a world with fewer jobs. The question for us is not how many roles will be lost, but what we will do instead. The future is watching.

This article is published as part of the Foundry Expert Contributor Network.
Want to join?