Looking Beyond the AI Haze 2: The AI Blind Spot

The AI story is being told through the rear-view mirror, driven by the latest announcement, warning or outbreak. Seeing through the haze means treating AI like the general-purpose technologies before it, and asking two questions: where does it genuinely fit today, and what will shape its path to 2030?

Few topics generate as much noise, and as little clarity, as AI. Will it make us less capable? Will it ever become a profitable industry? Could it destabilise the world economy, or escape the guardrails we set for it? For almost every finding there is a counter-finding, so the evidence alone will not settle these questions. Even the industry’s leaders do not appear to have a clear path forward.

The dominant narrative glosses over this. It says AI makes organisations more efficient and effective, and that this is broadly a good thing. The headlines tell a less settled story: AI increasingly cited in job cuts, growing fears of an investment bubble, lawsuits against chatbot makers, and the US, EU and China pulling apart on regulation. The narrative is papering over deep uncertainty, and it looks backwards, led by the technology. Technology may be running ahead of culture, but technology alone has never shaped the future.

So rather than starting with the technology, we start with history. AI is not the first general-purpose technology, and electricity, computing and the internet all have lessons to offer. Arvind Narayanan and Sayash Kapoor describe AI as a “normal technology”: volatile, but subject to the same dynamics as those that came before. That shifts the question from “What will AI do to society?” to “How will people, organisations and institutions incorporate AI?” It also gives us two lenses.

The first is product-market fit today. Bill Sharpe’s Technology Axes model holds that a technology only creates value when four things line up: the applications people demand; the systems, infrastructure and business models needed to deploy them; the social values and regulation that shape them; and the core technology itself. Even a toaster means little without standard voltages, a retail category and shared assumptions about what breakfast should be. But a toaster does one thing, and AI does not. Its general-purpose nature is part of what creates the haze, so the model works best applied to specific uses rather than to “AI” as a whole.

Fit today is only half the picture; the second lens is trajectory. Here, the variables that matter most sit largely outside the technology, and we group them as People, Power and Payoff. People: do workers embrace AI or resist it, and does public and political acceptance grow or harden into backlash? Power: does the market concentrate around a handful of providers, do regulators intervene, and does AI become a flashpoint in geopolitics? Payoff: does a viable business model emerge, or does spending keep outrunning returns, and is value widely shared or captured by a few? None of these will move in isolation, and how they intersect will produce markedly different futures by 2030.

This perspective forms part of Future Agenda’s work on The World in 2035.

Read more
Loading...

Foresights

Sort A-Z Sort A-Z

Previous

Next

${ item.value }

Previous

Next

Download PDF Of Foresight