Could artificial intelligence create new business risks?

Artificial intelligence is becoming a familiar feature of business software, but new research suggests that employers are increasingly concerned about the risks associated with its use.

Figures published by the Office for National Statistics on 5 November 2026 show that 12% of businesses with at least ten employees reported concerns about artificial intelligence affecting their supply chains during the following year.

This represented an increase from 5% in September and was the highest proportion recorded since the question was introduced in September 2024.

The findings relate specifically to supply chain concerns and should not be interpreted as evidence that 12% of all UK businesses have experienced problems caused by artificial intelligence.

Nevertheless, they provide an opportunity for business owners to consider how the technology is being introduced into their organisations.

Understand where AI is being used

Artificial intelligence is increasingly incorporated into accounting systems, customer management software, communications platforms and administrative applications.

Employees may also use separate AI services to draft correspondence, analyse information or prepare reports.

Although these tools can improve productivity, businesses need to understand what information is being processed and whether appropriate checks are in place.

For example, an employee using an AI service to summarise a confidential customer contract could inadvertently disclose commercially sensitive information if the service is not approved for that purpose.

Accuracy remains important

AI-generated information can appear convincing even where it contains mistakes or unsupported conclusions.

Businesses should therefore avoid relying on automated outputs for important financial, contractual or regulatory decisions without suitable human review.

The level of checking required will depend on the task. Drafting routine marketing material presents different risks from calculating payroll liabilities, interpreting legal obligations or recommending significant expenditure.

Introduce straightforward controls

A business does not necessarily need a lengthy policy before using AI, but it should establish some basic expectations.

Employees should know which applications are approved, what information may be entered into them and when the results must be checked by someone with appropriate knowledge.

It is also worth reviewing whether suppliers are introducing AI functionality into existing software and whether that creates new data protection, security or contractual considerations.

Measure the benefits

The commercial case for AI should be assessed in the same way as any other investment.

Businesses should identify the time or costs they expect to save, consider implementation and training requirements, and review whether the anticipated improvements are being achieved.

Used carefully, AI may help businesses improve productivity and reduce repetitive work. However, the greatest benefits are likely to come from combining new technology with clear procedures, appropriate oversight and a realistic understanding of its limitations. Source: ONS Business Insights and Conditions Survey, 5 November 2026.

Office for National Statistics