Ethics and regulation are bigger AI barriers than cost

Many SME leaders have a clear answer ready when asked why they have not adopted AI: it costs too much. It is a reasonable position. Budgets are tight, the technology is still maturing, and the ROI case can feel thin. But new UK government research published this week tells a more complicated story. Among businesses that raise concerns about AI, ethical worries rank as the single most significant barrier. Cost matters, but it is not first. For SME leaders and board members trying to understand what is actually holding their organisations back, that finding deserves a closer look.

Why the cost narrative took hold

The cost story spread because it is partly true and easy to communicate. Early enterprise AI projects carried enormous price tags. Media coverage of large-scale deployments reinforced the impression that serious AI required serious capital. For SMEs already cautious about technology spending, "we cannot afford it" became a shorthand that closed the conversation without requiring further explanation.

There is also a subtler reason the cost narrative persisted. Ethical and regulatory concerns are harder to articulate. Saying "I am worried about accountability if the model produces a biased output" demands more from a board conversation than saying "the licence fees are too high." Cost is concrete. Governance is abstract. The easier explanation tends to win.

What the evidence actually shows

The Department for Science, Innovation and Technology's latest AI adoption research gives a clearer picture. Lack of identified need and limited AI skills are the most commonly cited reasons businesses have not adopted AI. But among businesses that do raise concerns, the picture shifts significantly.

Eighty per cent of businesses that cite ethical concerns rate them as the most significant barrier to adoption. High costs come in at 76%. Unclear or uncertain regulation sits at 72%. Ethics outranks cost. That is not the story most SME leaders have been telling themselves.

The research also highlights that these concerns compound each other. A business uncertain about regulation is less likely to invest in building skills. A board uncomfortable with accountability questions is less likely to approve a budget. The barriers do not operate in isolation.

The ONS longitudinal data on AI in UK businesses points in the same direction. Adoption rates have grown steadily, but confidence in governance has not kept pace. Organisations that have deployed AI tools are increasingly reporting concern about whether their internal processes are sufficient to manage the risks those tools introduce.

Why this matters for SME leaders and board members

If your board is holding back on AI primarily because of budget concerns, but the real anxiety is about accountability, data security, and regulatory exposure, the two conversations will never align. You will keep presenting cost-reduction cases that do not address the actual objection.

Board members who raise governance concerns are not being obstructive. They are identifying genuine organisational risk. UK AI regulation is still evolving. The ICO has issued guidance on AI and data protection. Sector-specific rules are developing in financial services, healthcare, and legal. An SME that builds AI use cases without documented safeguards, clear ownership, and basic risk assessments is making a bet that those frameworks will not tighten.

That bet is getting harder to justify. The DSIT research signals that the UK government is paying close attention to how businesses are managing AI risk. SMEs that treat governance as an afterthought now may face costly retrospective work later.

There is also a people dimension that SME leaders often underestimate. AI adoption is roughly 80% a people and culture challenge and 20% a technology one. Team members who distrust how AI decisions are made, or who feel uncertain about their own accountability when using AI tools, will find quiet ways to avoid those tools. The technology sits unused. The investment produces nothing. Addressing ethical clarity early is not a compliance exercise. It is an adoption strategy.

A realistic approach to AI governance in your organisation

The good news is that basic AI governance does not require a legal team or a dedicated compliance function. SMEs can build meaningful foundations with a few structured steps.

Start by appointing an AI owner. This does not need to be a senior hire. It means nominating someone responsible for maintaining a simple log of where AI is being used, what data it touches, and who is accountable for its outputs. That log becomes the foundation for every other governance activity.

Document your use cases before you scale them. For each AI application, note the purpose, the data inputs, the decision it informs or automates, and the safeguard in place if it produces an error. This takes an hour for most use cases. It gives your board something concrete to review.

Monitor the regulatory landscape as part of normal business planning. The ICO publishes practical guidance on AI and data protection. The AI Safety Institute produces regular updates. Subscribing to these takes minutes. Ignoring them until a rule changes creates scramble.

If your team lacks confidence in these areas, structured training helps. gecco's Training and consultancy programme is designed for SMEs that need practical guidance on AI governance alongside the technical fundamentals. The programme addresses accountability frameworks, data stewardship, and compliance considerations in a format built for business professionals rather than legal specialists.

Honest limitations

It is worth being clear about what the DSIT research does and does not tell us. The data reflects businesses that cited concerns when surveyed. It does not capture the many businesses that have not engaged with AI at all and therefore have not yet formed specific concerns. The 80% figure for ethical concerns applies within the subset of businesses already thinking about adoption, not across the entire UK SME population.

This matters because the cost narrative may still be accurate for some organisations. A business that genuinely cannot identify a viable AI use case within its current budget is not making a mistake by deferring. The correction here is not "ignore cost" but rather "do not let cost be the only lens." If your board's unspoken concern is actually about governance, naming that concern and addressing it will move the conversation further than presenting a revised budget.

UK AI regulation is also genuinely uncertain in places. It is reasonable for SMEs to want clearer rules before committing significant resource. The practical response is not to wait for perfect clarity but to build governance structures flexible enough to adapt as rules solidify. Organisations that wait for certainty before acting on governance tend to be the least prepared when rules do arrive.

Find out where your organisation actually stands

If your organisation is weighing AI adoption and the real hesitation sits with governance, accountability, or regulatory uncertainty rather than budget alone, the AI Readiness survey is built to surface exactly those gaps. Take the AI Readiness survey. You will get access to 65+ free resources and a custom AI Readiness report. We then offer a free 45-minute AI Readiness call to walk through your results.

gecco's Training and consultancy service helps SME leaders and board members address the ethical and regulatory concerns that new UK government research identifies as the most significant barriers to AI adoption.

Website · LinkedIn · Case Studies · Newsletter