The singularity and our institutions: What must change?

Sam Altman’s recent claim that humanity has moved beyond the (event horizon) and that the technological take-off has begun was designed to sound historic. In his essay The Gentle Singularity, he argues that systems already exist that are smarter than people in many domains and that digital superintelligence is now within reach. Whether that judgement proves right is still contested. But the more useful question for governments and large organisations is not whether the singularity has formally arrived. It is whether the assumptions on which modern institutions were built would survive if it did. 
In the artificial-intelligence context, the singularity means a point at which machine intelligence begins to accelerate scientific discovery, software development and its own improvement so rapidly that human institutions can no longer reliably predict or absorb the pace of change. It is not simply a better chatbot or even an artificial general intelligence system. It is a shift from technological progress being managed by institutions to progress moving faster than those institutions can adapt.
That distinction matters because most public institutions and major companies were designed around scarcity: scarce expertise, slow information, periodic planning and hierarchical decision-making. Knowledge had to be gathered, reviewed and passed upward. Strategies were refreshed every three or five years. Regulation usually followed innovation. Authority rested partly on controlling information.
AI is already weakening those assumptions before any true singularity occurs. Stanford’s 2025 AI Index reported that 78 per cent of organisations used AI in at least one business function in 2024, up from 55 per cent a year earlier. Generative-AI use in at least one function rose from 33 per cent to 71 per cent. US private AI investment reached $109.1 billion, compared with $9.3 billion in China and $4.5 billion in the UK. 
The economic upside is potentially enormous. PwC estimates that broad AI adoption could raise global output by as much as 15 percentage points by 2035, effectively adding about one percentage point to annual growth. Yet the same research stresses that capability alone is insufficient; deployment, trust and responsible use will determine how much value is actually realised. 
This exposes the first central problem: institutional inertia. Technology can spread in months, while organisations may require years to redesign workflows, amend regulation, retrain staff and reallocate budgets. McKinsey found that 78 per cent of surveyed organisations were using AI in 2024, but only 21 per cent of those using generative AI had fundamentally redesigned at least some workflows. In its 2025 survey, nearly two-thirds still had not begun scaling AI across the enterprise, and only 39 per cent reported an enterprise-level EBIT impact. 
The gap is not between organisations that own AI and those that do not. It is between those that use AI to automate yesterday’s processes and those willing to question whether those processes should exist at all.
If Altman’s forecast proves broadly correct, the benefits could be transformative: faster medical discovery, more productive firms, more accurate forecasting, personalised public services and cheaper access to expertise. Knowledge scarcity could decline sharply. Smaller economies might gain capabilities once available only to large states or multinational corporations.
But the risks would be equally structural. Labour markets could adjust faster than education systems. Decision-making could become more concentrated in institutions controlling models, data and computing power. Errors could scale at machine speed. Cyber threats, disinformation and regulatory arbitrage could outrun existing safeguards. The youngest workers may feel the disruption first: Stanford’s 2026 AI Index reports that employment among software developers aged 22 to 25 fell by nearly 20 per cent from 2024 in AI-exposed segments. 
The deeper danger, however, is not that machines become smarter than people. It is that institutions remain slower than the environments they are expected to govern.
The required response is therefore not another technology programme. Institutions need to review the assumptions embedded in their operating models: where authority sits, how evidence reaches decision-makers, how frequently strategy is revised, how quickly regulations can be updated and which workflows AI makes unnecessary. They must also separate decisions that can be automated from those requiring political accountability, ethical judgement and human legitimacy.
The singularity may arrive later than Altman expects, or not in its strongest form. That does not weaken the case for reform. Institutions that become more adaptive, data-literate and capable of continuous learning will gain even under a gradual AI scenario. Those that simply digitise existing bureaucracy may discover that technology has accelerated everything except the institution itself.
The defining competition of the next decade may therefore be less about which country or company possesses the most powerful model. It may be about which institutions are prepared to abandon assumptions that no longer hold, learn continuously and govern at the speed of change.

Dr Ziad Alzaidi
Strategic Management & Organisational Transformation Consultant, UK / Oman