AI Values Institute warns on AI value alignment gap

The AI Values Institute has published a white paper on what it calls the AI Value Alignment Gap, setting out a framework to assess whether the value created by artificial intelligence matches an organisation's stated principles.

The paper argues that as AI adoption spreads across sectors, many organisations have focused on efficiency, cost reduction and productivity. Boards and executive teams are under pressure to move quickly and show measurable returns, but often lack a clear way to judge who benefits from AI-driven gains and who bears the costs.

That question sits at the centre of the Institute's framework. The report defines the AI Value Alignment Gap as the distance between an organisation's declared values and the outcomes generated by its AI systems.

According to the paper, the gap can widen when companies scale AI tools faster than they can assess broader effects on staff, customers, communities and society. Operational measures such as speed, efficiency and financial return are often treated as the main indicators of success, while trust, wellbeing, fairness, inclusion and long-term resilience receive less attention.

Broader impact

The argument comes as companies face growing scrutiny over how AI is used in business operations and decision-making. While governance and compliance have become established areas of focus, the report contends that they do not fully address how AI-created value is distributed or whether that distribution reflects human-centred values.

The paper highlights several challenges: a tendency to measure AI outputs rather than outcomes, an uneven distribution of value across stakeholders, and governance systems that focus more on risk than on value creation and distribution. It also argues that productivity gains can bring unintended human, cultural and social costs.

Edosa Odaro, Founder, AI Values Institute, said the central issue for organisations has shifted beyond whether AI can deliver returns.

"The challenge facing organisations is no longer whether AI can create value. The challenge is understanding what kind of value is being created, who receives it, who bears the cost, and whether those outcomes are aligned with the values leaders ultimately stand for," said Edosa Odaro, Founder, AI Values Institute.

The Institute presents the paper as a practical tool for leaders who need clearer language to discuss trade-offs in AI deployment. Many organisations do not yet have a framework that links technical decisions to human outcomes in a structured way, it says.

Leadership pressure

The report argues that this matters because AI has intensified competition among businesses seeking technological advantage. In that environment, leaders are expected to deliver innovation and performance while also addressing rising expectations around transparency, accountability and social impact.

Its core claim is that performance alone is becoming an inadequate measure of success. Organisations seeking long-term legitimacy will need to test not only whether AI systems work efficiently, but whether their effects support the values they claim to uphold, the paper argues.

Odaro said those effects can vary sharply depending on a person's position within an organisation or system.

"AI does not create value in isolation. Value is experienced differently depending on where you sit within a system. What appears as efficiency for one stakeholder can feel like loss for another. The organisations that succeed will be those that understand these dynamics and actively design for shared value creation," said Odaro.

The publication forms part of the Institute's broader work on AI stewardship, focused on research, education and practical frameworks for decision-makers. It argues that leaders need tools that go beyond implementation questions and compliance checklists so they can ask harder questions about impact, responsibility and trade-offs as AI becomes more embedded in strategy and operations.

The paper concludes that closing the AI Value Alignment Gap is likely to become a defining leadership issue in the AI era, as organisations are pushed to examine not just what AI can do, but what it should do, for whom, and at what cost.