As companies race to integrate AI into their operations, the discussion around governance has stumbled into the wrong place. Regulators are deliberating over mandates, policymakers are discussing guardrails, and developers are arguing about technical controls. These questions are important, but they ignore the most urgent driver of responsible AI governance: the people who use these systems every day. Without investing in workforce capabilities, organizations risk building harm into their operations and finding themselves liable when things go wrong.
Adoption of AI does not wait for governance to catch up
Companies are incorporating AI tools wherever they can to achieve efficiency and revenue gains, with or without oversight frameworks in place. Recent news from the UK illustrates this tension between governance and innovation. In the same week that the Treasury Committee warned against the financial sector’s adoption of artificial intelligence in particular Risk of causing “serious harm” For society and the economy, Lloyds Banking Group announced that the adoption of artificial intelligence Increase its revenues in 2025 By 50 million pounds ($66.8 million).
So the governance risk is not just that AI is advancing rapidly. Here, risks also stem from the fact that AI is being integrated into workplaces where employees are not equipped to understand its limitations, failures or compliance implications. This gap is where new governance concerns emerge.
Governance risks of deploying AI without literacy
The most predictable outcome of poor AI adoption is what practitioners call “shadow AI.” Without formal training, employees resort to unapproved consumer-grade tools to complete professional tasks, often without detection. in the united kingdom, 81 percent of AI users Do not disclose the use of AI to managers. Sensitive company data can be entered into generic forms that retain the inputs or reuse them for further training, creating new regulatory and reputational risks.
The problem is exacerbated when employees misunderstand how AI actually works. Employees may treat AI as a fact-based search engine rather than a pattern-based reasoning engine, and fail to critically evaluate the accuracy of its output. Take, for example, the widely reported cases of lawyers being sanctioned Introducing “hallucinations” generated by artificial intelligence In court files. When users cannot effectively evaluate AI outputs, it is the employer who bears responsibility, undermining trust with customers and regulators.
Bias represents another limit to governance. AI systems inherit patterns from their training data. If employees fail to recognize discriminatory outcomes, they risk embedding systemic bias into operational decisions. In 2021, this issue came to the fore in the United States with the report that it was discovered to be automated Lending systems rejected up to 80 percent of mortgage applications from black applicants. Similar failures have since appeared in the algorithms used Evaluation of social welfare applications and Job applications. From a governance perspective, this creates significant ethical, legal and reputational risks, not to mention broader impacts on human rights and social justice.
Even in cases where harms do not materialize, the deployment of low-skilled labor limits the return on investment. Technology rollout is not synonymous with digital transformation. Without redesigned workflows and trained staff, AI produces piecemeal productivity gains rather than company-wide impact.
Building governance from the ground up
In Europe, the workforce dimension of governance is already being recognized. the EU law on artificial intelligence Incorporates knowledge of AI as a legal requirement for employees handling AI systems. In the absence of similar regulation in the United States, companies must lead these efforts themselves. Based on our experience advising organizations on AI governance, A A reliable bottom-up approach It is based on three interconnected foundations.
The first is AI literacy, which varies by role. For executives, literacy means knowing what questions to ask: How do we monitor for bias? Who is responsible for the model’s performance? When does human review overtake AI output? Leaders must be able to evaluate whether AI is a strategically appropriate response to a business problem, rather than an ad hoc response.
For technical teams, AI literacy means responsible data management, model validation, performance monitoring and documentation. For end users in other roles, such as recruiters using AI screening tools, marketers crafting AI-powered campaigns or analysts using generative AI as research assistants, literacy is practical and procedural. It includes understanding the tools adopted, verifying the outputs, knowing how to escalate concerns, and applying human judgment.
The organizations we have worked with that are ahead of the curve differentiate literacy training by role, treating it as an operational skill linked to accountability.
The second basis is updating policies and procedures. Acceptable and clear usage policies reduce the potential for AI shadowing, prevent over-reliance on outputs and demonstrate accountability for AI-supported decisions.
Policies governing AI supply chains and procurement require scrutiny. AI vendors must undergo structured due diligence covering training, data management, bias mitigation processes, monitoring capabilities, and contractual clarity around liability. As we wrote Within the framework of institutional sustainabilityEven well-intentioned organizations can undermine their governance efforts by relying on a poorly audited supply chain.
The third pillar is clear accountability structures across the AI lifecycle. This may include cross-functional AI governance committees, responsible AI leaders, risk oversight at board level or the involvement of independent assurance providers. The structure will vary depending on the size of the organization and sector. What matters is that accountability is clear, and that governance is integrated into product development, procurement, compliance, and risk management rather than treated as a separate practice.
Responsible governance of AI as an investment, not a liability
Discussions on AI governance will continue at the regulatory level. Standards will evolve, and implementation landscapes will change. Many of these factors remain beyond the control of any single company. Manpower capacity no.
Reframing AI governance around employee investment, updated policies and clear accountability returns agency to business leaders. It also provides a constructive counterweight to concerns about job displacement due to AI: rather than replacing workers, responsible AI governance equips and upskills them. Those organizations that take this seriously will be better positioned to maintain trust with customers, regulators and the public as scrutiny of AI adoption continues to increase.
Amelia Williams He is the Chief Research Impact Officer at Triliteral Research and has experience communicating science at the intersection of emerging technologies, environmental issues, ethics, and policy. At Tritraler, you support the development and implementation of research projects alongside policy, media and industry engagement.
