Britain Is Embracing AI — But Are We Ready for What Comes Next?-

22nd August 2026

Britain is adopting artificial intelligence at remarkable speed.

But there is an intriguing problem developing beneath the headline figures.

Businesses are increasingly embracing AI while the public appears to be becoming more wary of it.

That may prove to be one of the most important tensions surrounding the technology over the next few years.

The latest Office for National Statistics figures show that the proportion of UK businesses with 10 or more employees reporting that they use at least one form of artificial intelligence has risen from around 12% in late 2023 to around 35% by June 2026. In less than three years, business adoption has almost tripled.

That sounds like the beginning of a technological revolution.

And it may well be.

But another ONS survey, conducted among the public in June, suggests that many people are looking at that revolution rather differently.

Only 13% of adults said they believed AI brought more benefits than risks, while 38% said it brought more risks than benefits. A further 43% thought the benefits and risks were about equal.

The proportion believing the risks outweigh the benefits has risen sharply from 25% in August 2024 to 38% today.

Britain is therefore entering an extraordinary period in which the technology is spreading rapidly even as confidence in its consequences appears to be weakening.

The business revolution is real — but still relatively shallow

The ONS business figures need some qualification.

Although 35% of businesses with 10 or more employees report using AI, the depth of that adoption remains limited.

The average number of AI technologies being used by an adopting business has risen only from around 1.4 to 1.6 since late 2023.

Only 10% of businesses that use AI describe their use as extensive, while just 15% say that more than half of their employees use AI as part of their daily work.

So we are not yet looking at companies being transformed from top to bottom by artificial intelligence.

In many cases, AI is being introduced one task at a time.

Someone uses it to draft correspondence.

Another employee uses it to summarise documents.

A marketing department uses it to produce images or ideas.

A business owner uses it to research a subject or analyse information.

That may sound modest.

But multiply those small changes across thousands of businesses and the potential economic effect becomes much more significant.

The technology is spreading faster than the transformation

This is perhaps the most important distinction in the ONS report.

AI adoption is accelerating, but deep integration has not yet caught up.

Large language models are the most widely used AI technology among businesses, followed by visual-content creation and machine-learning applications for data processing. Robotics, interestingly, remains much less common.

This suggests that Britain's AI revolution is currently much more about information and knowledge work than about robots replacing people on factory floors.

And that may explain why some of the more dramatic predictions about mass unemployment have not yet materialised.

AI is changing jobs before it necessarily eliminates them.

An administrator might spend less time producing routine documents.

A designer may use AI to develop initial concepts.

A researcher may use it to sift through information.

A small business owner may suddenly have access to assistance that previously required another employee or an outside consultant.

The job remains.

The work changes.

But the public is not necessarily convinced

This is where the second ONS report changes the story.

When asked whether AI would benefit them personally, people were far from universally enthusiastic.

The public's concerns extend well beyond employment.

A particularly striking 81% of adults were concerned that AI could make it harder to tell whether news or information was fake, while 77% expressed concern about their personal data being used without their consent.

These concerns are not trivial.

People are increasingly being asked to live in an environment where seeing something is no longer necessarily evidence that it happened.

A photograph can be generated.

A voice can be replicated.

A video can be fabricated.

An apparently authoritative piece of writing can be produced in seconds.

For democracy, journalism and everyday life, that creates a problem which goes far beyond whether AI can save a business a few hours of administrative work.

And then there is the jobs question

The employment issue remains particularly sensitive.

Younger adults are considerably more likely to have already used AI for work or education. Among 16 to 29-year-olds, 63% reported using AI for work or education, compared with only 3% of those aged 70 and over.

That isn't particularly surprising.

But it produces an interesting contradiction.

The generations most familiar with AI are also confronting the possibility that it could alter the jobs they are preparing for.

That doesn't necessarily mean mass unemployment.

It could instead mean that the definition of a job changes.

The graduate entering an office in five years may be expected to work alongside AI from their first day.

The skill employers value may become less about producing a first draft and more about knowing whether the machine's first draft is any good.

That is a profound change.

There is already a gap between employees and employers

Another intriguing finding is that AI may actually be spreading through workplaces faster than businesses realise.

The ONS business survey found that around 35% of businesses with 10 or more employees reported using AI.

But the ONS's wider analysis found that more than half of employees reported using AI, although this measure includes AI use for work or education and can therefore include activity outside formal business systems.

That suggests a kind of unofficial AI revolution is taking place.

Employees can experiment with AI without waiting for their employer to introduce a formal AI strategy.

Someone discovers that a task taking two hours can be completed in 20 minutes.

They start using the technology.

Productivity improves.

But perhaps nobody has checked what information is being entered into the system.

Nobody has established an AI policy.

Nobody has considered whether the result is accurate.

The technology has entered the workplace from the bottom up.

That creates an enormous opportunity for small businesses

This is where I think AI could become particularly interesting for Scotland and places such as Caithness.

A small business cannot afford to employ a specialist in everything.

It may not have a marketing department, data analyst, graphic designer, researcher and IT specialist.

AI potentially gives a single business access to capabilities that would previously have been beyond its budget.

And geography matters less.

A small company in Caithness can access many of the same AI tools as a company in Edinburgh, Manchester or London.

That doesn't eliminate the disadvantages of living and doing business in a remote area.

But it could reduce some of them.

For a small Scottish business, AI may therefore represent something more interesting than automation.

It could be a way of narrowing the gap between small and large companies.

But there is a danger in believing the machine

The ONS findings also suggest why businesses should be cautious.

AI can produce useful answers.

It can also produce convincing nonsense.

It can misunderstand a question, invent information or confidently present something that is simply wrong.

The more businesses depend upon it, the more important human checking becomes.

There is also the issue of confidential information.

A business that casually feeds customer details, commercially sensitive documents or financial information into an AI system may create risks it has not considered.

The temptation will be to assume that because the technology is easy to use, it is also safe to use.

Those are very different things.

Britain could therefore have an AI trust problem

This may ultimately become just as important as the technology itself.

Businesses are adopting AI because they can see potential gains in productivity and efficiency.

Consumers, meanwhile, are asking whether they can trust the technology.

Can they trust an AI-generated news story?

Can they trust an image?

Can they trust an online review?

Can they trust that their personal information is safe?

Can they trust that an AI-assisted decision about them has been made fairly?

And, increasingly, can they trust that their job will still exist in its current form?

The ONS figures show that these questions are not theoretical.

Public concern is already rising.

The paradox of the AI revolution

Perhaps the strangest thing about artificial intelligence is that it may become economically important before people become comfortable with it.

Businesses don't necessarily need everyone to love AI.

They need it to work.

If an AI system can save a company thousands of pounds, reduce administrative work or allow a small firm to compete with a much larger one, the economic incentive to adopt it will remain powerful.

That means public hesitation may not slow adoption as much as some might expect.

Instead, it could create pressure for better regulation, greater transparency and clearer rules about how AI is used.

We may be approaching the second phase

The first phase of Britain's AI experiment has been relatively simple.

Can we use it?

The answer is increasingly yes.

The next question is much harder.

What happens when we reorganise businesses around it?

That is when the consequences for productivity, employment, wages and competitiveness could become much more significant.

The ONS figures suggest that Britain has already crossed the threshold into widespread experimentation.

But the relatively shallow level of adoption suggests that the biggest economic effects may still be ahead.

At the same time, the public is telling us something that businesses cannot afford to ignore.

Technology may be advancing rapidly.

Trust isn't.

And ultimately, the success of the AI revolution may depend on bringing those two things together.

Britain may well discover that the hardest part of artificial intelligence isn't teaching machines to become smarter.

It is teaching people to decide when they should believe them.