Acknowledge uncertainty but communicate clearly about the coming upheavals from AI
AI is no longer quietly rewriting workflows. The alarms are coming from inside the boardroom. Executives at Ford, Amazon and JP Morgan are publicly warning of deep job losses as AI reshapes their operating models. Ford’s Jim Farley has said half of all white-collar roles are under threat. McKinsey is rattled saying ‘this is existential’. Amazon’s Andy Jassy calls AI the next wave of disruption. JP Morgan has already accelerated large-scale automation behind closed doors. These are not forecasts from think tanks, they are strategic signals from the people closest to the machine, and it’s fair to say, maybe with a thumb on the scale.
Recent US labour data supports the shift, although admittedly with tariffs and other Trumponomic madness there are certainly other factors at play too. A US government analysis published in July estimated that three per cent of American workers have already been replaced by AI tools, with the trend accelerating fastest in professional services and administration. The World Economic Forum anticipates up to 83m jobs eliminated globally by 2030. Forecasts vary, but momentum is clear. AI is replacing tasks and shrinking teams as much as streamlining workflows. Self-actualisation and UBI is not around the corner for most of us.
So what’s next?
Executives must be honest about what is known and what remains unknowable. McKinsey and other analysts forecast that full AI adoption across industries may stretch into the 2040s. However, most predictions underestimate the compounding impact of self-improving tools. AI models are not only faster and cheaper, they also learn how to automate new roles without human prompting. Generative AI’s capability to write code, summarise legal briefs, interpret medical data and manage customer relationships is growing weekly along with troubling ‘emergent abilities’. Once a single process is automated, follow-on impacts cascade through adjacent roles. The result is not linear disruption, it’s exponential.
No one can say with certainty how many AI engineers or prompt supervisors the world will need in three years – my bet is less than people would like. Planning on historical timelines will not help. Past industrial revolutions were physical. AI is cognitive and scalable even with massive infrastructure needs. Shifts can, and are, happening inside browser windows. The analogy is not to electricity or the printing press. AI is a billion lightbulb moments all happening at once, some seismic, some infinitesimal that will have gargantuan knock-on effects. We’ve never had the perfect storm of infrastructure, distribution, choppy social waters, geo-political uncertainty making folks jumpy and greedy.
AI is a billion lightbulb moments all happening at once, some seismic, some infinitesimal that will have gargantuan knock-on effects
Executives need to acknowledge that generative models introduce something no previous technology could: unpredictability. AI tools can hallucinate, deceive or operate outside intended scope. And that’s right now. What’s allowed to happen in the future, appears to be less and less likely to be policed if current policies are adopted. Engineers now spend as much time supervising and correcting outputs as generating them. Productivity gains are real, but so are novel forms of risk. Redundancies are not just about efficiency, they are also about trust, liability and system control. Right now, a lot of CEO’s are bragging that their once bloated companies will be streamlined. These lovely folks are missing the point. We’re about to have a lot more poor people that can’t afford their things.
Employees see what is coming. Silence from leadership is a risk in itself. Even without a full plan, managers must start talking, and clearly. Honest communication that acknowledges uncertainty builds credibility, so start talking. Waiting for clarity before engaging staff delays adaptation and amplifies anxiety, and workers aren’t expecting perfect answers, but they will be expecting respect and context. Current advisors (myself included) are being inundated with calls to not just explain the technology and potential, but also to help shape a coherent internal narrative so as not to freak everyone out, and likely cause some folks to jump.
Reframe the next era
Sam Altman himself warned in June that AI will cause job losses and national security threats, and he’s not alone. Executives building the future are more vocal about the social impact than many governments. Companies who dismiss these warnings as hype or worst-case posturing miss the strategic opportunity. Businesses still have time to shape how AI augments their workforce. Those that treat this shift as an HR issue will lose ground to those who reframe it as a strategic transformation. Now more than ever, my favourite phrase seems apt; disruption is an invitation.
Restructuring work is only part of the puzzle. Three major new models are about to drop that will redraw the roadmap again. OpenAI’s GPT-5 (end of August), Grok 5 from Elon (likely December – four just dropped), and DeepSeek’s R2 reasoning model (any day), each promising another leap in capability, scale and unpredictability. These systems won’t just accelerate automation, they will trigger strategic resets. Reasoning, planning and real-world integration are the next frontier, and any business waiting for certainty before responding is already late. You’ve got to be making moves now.
Executives need to build adaptive strategies that assume shifting foundations. Plan for job removal, not just augmentation. Model impact scenarios across operations, customer service, design, compliance. Not every role disappears, but every role changes. Secondary effects will be brutal. Reduced employment means lower consumer spending. Dislocated workers will bring reputational, political and economic backlash. Governments need new plans for income security, training and digital infrastructure. No current policy is likely close to ready, and no administration is ready for the upheaval.
The next 6-12 months
In the next six to twelve months, businesses should focus on three actions: audit every repeatable process for automation potential, build internal AI fluency beyond tooling to include critical oversight, and create a standing cross-functional team tasked with exploring the known unknowns. Strategic foresight needs to become operational muscle. Legal, HR, data, product and C-suite leaders need to collaborate now, not once the layoffs begin.
A good start is creating a procurement kill list. Audit all current tech vendors and platforms for AI-readiness and long-term viability. Flag contracts that assume manual workflows, offer no roadmap for intelligent automation, or resist integration. Begin replacing those vendors now, not once legacy processes become the bottleneck.
Plugging the gaps in understanding is not optional. These systems are fast, opaque, and increasingly agentic. Where humans used to be gatekeepers, we now get models that can decide what to do next. Currently with our permission but for how long until that is written out by humans or machines. The issue is not just jobs, it’s the redefinition of agency and accountability inside organisations. The smartest move any executive can make right now is to build a learning agenda, engage with outside voices who see across industries, and be unflinchingly honest about what might come next.
Inviting disruption so you explore it, before it hits is a smart move, Take the invitation and treat it as a gift. Every business gets to choose whether to cross the threshold prepared, or be caught in the undertow of tools it doesn’t yet understand.
Paul Armstrong is founder of TBD Group and author of Disruptive Technologies