False Certainty

Future fiction presented as future fact

Photo by Abid Shah on Unsplash

“The UK stands to gain £40 billion per year in public-sector productivity improvements by embracing AI, amounting to £200 billion over a five-year forecast.” said the Tony Blair Institute for Global Change in 2024. When I first read this, I probably rolled my eyes. Not because I know it’s false - but because I can’t make sense of it, and yet it’s presented with confidence and certainty. The numbers felt a bit… wild.

Digging on these numbers - that £200m figure appears to be grounded on an assumption that one-third of the entire UK public sector workforce (540,000 people) have been made redundant by 2029 and replaced by AI at 1/20th the cost¹. That’s very clearly absurd, even for 2024 when AI² was more ripe with potential, and less road-tested as it is today.

In defence of the paper, its admirably structured argument and referencing shows the inner workings, and has allowed me to make the point. This example is not unique. It’s false certainty, and you’ll see it everywhere when you start to look.

The predictability limit

We grapple with fact checking the past, but too often, we give the future a free pass. Too often, future fiction is presented as future fact. The more factual the presentation of the future, the more grounded it must be - evidence-based, and a proven method of prediction. Equally, there must be room for creativity and subjectivity, so long as this is honestly presented for what it is.

There’s a limit to grounded predictions - the predictability limit. Beyond this point, creating the appearance of fact is to create false certainty.

Here’s a simple explanation of this principle - the weather forecast. Predicting the maximum daily air temperature in three days’ time, to single-degree accuracy is grounded - decades of science and modelling mean that in much of the world, the forecast is consistently within 3-4 degrees of later recorded temperature. That’s a useful range for most people, and most people wouldn’t be impacted if they make plans based on the forecast. The sense of certainty isn’t misleading for most. It’ll be hot at the weekend - so let’s make a plan to go to the beach.

Predicting the air temperature in 20 days’ time is only accurate to within 10-12 degrees. This variability means the forecast isn’t very useful for most people. It’s hard to make plans where the weather really matters. If weather forecasters tell you the weather is specifically 23℃ in 20 days’ time, this is misleading and unhelpful. In the UK, given our changeable climate, you probably shouldn’t be planning a day sunbathing on the beach 20 days ahead.

These misleadingly specific predictions are beyond the ‘predictability limit’ for weather forecasting - which, with today’s technology and science, is around 10-14 days. Domains where human systems, societies and economies are part of the mix, prediction is more chaotic, but there’s a still a predictability limit. The limit is harder to see, maybe more subjectively located - but cross it by a distance and the false certainty will become apparent.

What do weather forecasters do beyond the predictability limit? They change how they communicate. mid-July’s weather, in 15 days’ time, is “High pressure patterns are looking slightly more likely than low pressure patterns during the early part of this period, meaning a greater chance of settled and drier than average conditions.” The probabilities in the language show the uncertainty - the forecast remains an honest representation.

False certainty lies in being overly specific about a point too far into the future. This is done by stretching a grounded methodology for prediction with excessive assumptions.

Ultimately, it’s making stuff up about the future, whilst pretending you’re not.

False certainty is everywhere

In the wider world, false certainty is so common, you’ve probably stopped noticing it.

It’s in our politics: such as the idea that £350m-a-week saved from EU spending would contribute to a post-brexit NHS or HMT’s analysis which concludes Brexit will leave households “£4,300 worse off in 2030”. Both these claims skim over several years of macro-economic uncertainty, the uncertainty of the period and terms of a negotiated Brexit deal, the uncertainty of government spending and policy choices, to make specific predictions. They exceed the predictability limit in years without changing the language.

It’s common in the world of consultancy as a way of gaining attention, such as the claim, in syndicated media, cited to Gartner in 2022 that “25% of people will spend at least one hour per day in the Metaverse by 2026”. It’s now 2026, they obviously don’t - but there will have been analysis behind this claim with outlandish and opaque, assumptions. The metaverse has no meaningful presence today, beyond a rebranding of tech that pre-existed the metaverse hype. Gartner exceed the predictability limit of perhaps just a few weeks after it was written.

Inside organisations, the picture is mixed. There are organisations that avoid false certainty well, and their internal communication and collaboration values ways of understanding uncertainty and using it to learn and make decisions.

But inside many large, established institutions, false certainty is commonplace. Strategies, business cases, plans and roadmaps all fall foul of the need to sell to the system - and to make the sale, they use false certainty. They tell the story that needs to be told - of value delivered, risks mitigated, or efficiencies achieved. Because understanding uncertainty is less valued, it’s less communicated. Home truths about uncertainty are reserved to informal, more private spaces between trusted colleagues.

A compelling example of this is the UK government’s portfolio of major projects. 15%, or total value of £198 billion of projects are ‘red’ rated, indicating they can’t deliver what they promised to deliver. This is a snapshot - with projects typically drifting from green to red over time - so many of the other 85% of projects have red in their future. The gradual reveal of a mismatch in expectation is the pattern, not the exception. HS2 is by far the most prominent example from within this portfolio, beginning life as a £15 billion idea, approved for £37 billion, with current cost forecasts disputed, but ranging up to £170 billion.

A judgement of when you’ve crossed the line from honest storytelling into false certainty is subjective. Sometimes expertise is needed to get a feel for whether there are too many assumptions - but often just hearing the method behind the numbers is enough to raise an eyebrow. Arguably, a little false certainty is acceptable - the narrative may serve a greater good, the simplicity of the argument may make communication more effective, the sense of confidence that certainty creates may win hearts and minds for a challenge ahead.

False certainty is widespread because there’s an advantage gained from getting people to believe in it - you have more buy-in, more backing, more support for your argument. If anyone does think you’re making stuff up, it’s still a hard job to show the falseness in false certainty - particularly if you don’t share your working.

The allure of false certainty

False certainty is alluring for those trying to make an argument. More specific, quantitative information brings a pop and sparkle to storytelling. Within organisations, false certainty is a reliable tool for obtaining funding, headcount, and senior level or investor backing.

But if the gap between reality and false certainty becomes too wide, it brings a long shadow of disappointment, mistrust, and cynicism. In democracies, it erodes our trust in institutions, in politicians, in incumbent governments. Trust in politicians is at a historic low, in part fuelled by a cycle of over-promising and under-delivering.

In organisations, it erodes optimism and opens fracture lines where reality meets the falsehoods. Many have experienced the long shadow that falls over a failing project, as reality begins to expose the widening gap, whilst promises and expectations stay fixed. It’s hard to work in those conditions, and it can encourage toxic behaviour to emerge. For people delivering and operating, it’s that growing feeling that the work you do is disconnected to what leaders are saying. For leaders it’s the anxiety when milestones are missed and outcomes fail to materialise, but you don’t really know why.

Customers and stakeholders also experience the long shadow through the tangible impact on services and relationships. Delivery and operations are simply less effective when the workforce is chasing after false certainty, rather than learning what really works, and adapting the plans and strategies based on evidence.

What’s the alternative

I’ll focus on what I know first-hand: institutions, private and public. (I think there are lessons which may apply more broadly, but the challenges of false certainty in politics, media and consultancy are more deeply rooted.)

These are a few institutional ingredients which I’ve seen disincentivise false certainty.

Find a way to show the uncertainty

It won’t be easy, but it’s important to show the uncertainty of your work - especially if no-one’s asking for it.

When everything feels uncertain, people often say nothing at all about the future. This is really common at the early stages of iterative and incremental delivery - you feel you know the least, so forecasting feels at its most dangerous. The problem is that the gap will be filled by whatever narrative suits others - perhaps a falsely certain narrative, perhaps doubt and fear.

Communicating uncertainty is hard. It boils down to things like probabilities, ranges of cost, or time. It’s “we’ll deliver sometime between 2027 and 2030” which regardless of how much you show the truth of this uncertainty, most organisations will struggle to accept it. So the greatest challenge is convincing people that uncertainty is grounded in evidence, not ineptitude or lack of confidence. Showing this uncertainty can help to bootstrap more mature dialogue - if some scenarios are clearly unacceptable, then it’s clear a new strategy is needed.

Communicating uncertainty is helpful because it better represents reality. But humans find it hard to accept or understand. Our innate biases mean we struggle to understand and conceptualise probabilities. There are ways around this - such as telling better stories of how risks emerge and impact - connecting people to what-ifs, that they otherwise can’t see or feel.

If the communication challenge can be overcome, there’s enormous reward. Communicating uncertainty can be an incredible enabler of test and learn methods - the ability to learn from delivery, and adapt your strategy continuously. Ultimately, is a necessary condition of good design, and really meeting the needs of users. So if you want your government services to be good, and you want a great experience when you next buy that TV or takeaway, you should hope those institutions know how to communicate uncertainty.

Show intent through outcomes

When leaders show their intent through the outcomes they want to see: “I want the steady growth of customers, but more importantly to keep our place as the most trusted product in the market.” Shared outcomes help create a culture where false certainty doesn’t get you ahead. Instead, what helps teams succeed is transparency, nuanced dialogue between trusted peers, making trade-offs in collaboration.

With outcomes, leaders can still be strategically clear, but they’re open to hearing ideas about how outcomes can be achieved. Enabling more creativity in strategy and plans that can keep changing as new information is learnt. Leaders’ trust remains, so long as there’s a continuous push towards the outcome and transparency about progress.

Hypotheses

Hypotheses are a natural direct alternative to false certainty. They’re a technique common in science, popular for individual teams using agile methodologies or product thinking, but less common outside these domains - and crucially, rare in the boardroom.

Hypotheses have an if… then… because… structure. For example: “If we introduce AI tool X, then we expect our workforce to be 20% more productive, because they’ll spend less time reading complicated guidance documents and more time providing answers to customers.”.

This structure negates false certainty because it doesn’t present bold strategic leadership as fact, but as a falsifiable assertion. It allows leaders to be more proactive than simply stating outcomes and leaving the rest to their teams. They’re a natural extension of outcome-based working.

It makes it safe to challenge the approach, if there’s evidence it’s not having the intended impact. The structure requires an explanation to the strategy - the mechanism by which the impact happens. Again, if this mechanism fails to materialise, the hypothesis can be challenged. However, it does allow leaders to be clear, show a more singular direction, keep their strategic communication simple, and avoid the language of doubt which can undermine confidence.

Price in risk

Put risk into your estimates. Don’t put them off to the side, and tell stories about the happy path to delivery. Put risks inside your delivery story. Use risk to unpack scenarios that widen the range of possible futures.

Make risk a collective ownership challenge - everyone’s problem, not specific to an ring-fenced area of delivery like a project. For example, don’t seek contingency of time and budget. Instead, anticipate the risk, and have a backup plan for where the time, money and skills will come from to handle the risk.

Doing this well is hard work

These ideas aren’t quick fixes for organisations where they rarely happen. They’re deeply transformative for many organisations. To mitigate false certainty in an organisation where it’s common already, will be a process which takes years - not a specific number of years - but quite a few 😉.

But it can begin almost immediately, where leaders can see the value of avoiding false certainty, and change their language and behaviours to reflect this. Teams can do the same in the spaces they influence - and communicate uncertainty upwards to anyone who will listen.

---

It’s not an accident I chose the weather analogy to explain my point. I’m writing this piece during the European heatwave of June 2026. I’m looking forward to a much cooler July and August, but I honestly don’t know if it’ll come.

1. The original source of this is an unpublished government analysis by CDDO which an NAO report caveats with “[CDDO] did not examine the feasibility of delivering these productivity gains”. Neither the NAO nor CDDO implied a 5-year timetable, which was added in the TBI analysis.

2. I’m consciously using AI here as shorthand (as many do) for the mainstream impact we’ve seen since 2022, post ChatGPT release, and most closely associated with Large Language Models.