Artificial intelligence is often discussed as a race defined by scale, speed and computational power. The more important question is whether these systems are shaping a future that strengthens human agency or quietly erodes it.
The conversation around artificial intelligence has become increasingly technical. Headlines focus on model size, benchmark performance and competitive advantage, creating the impression that the story of AI is primarily an engineering contest. Yet beneath the noise sits a more uncomfortable question, one that executives cannot avoid indefinitely, what value is being created for people, communities and society itself.
Edosa Odaro, Senior Executive Advisor for AI and Data, believes that the industry’s obsession with technical progress has obscured the real objective. His argument challenges the prevailing narrative. Success, he suggests, should not be measured by accuracy scores or processing capability, but by whether AI improves human outcomes in ways that can be felt beyond the technology team.
“I have been in countless boardrooms where executives celebrate ninety per cent accuracy in models while customers are literally walking out of the door,” Odaro says. “Companies are spending billions on AI and claiming huge productivity gains, yet when you ask them to correlate that with customer satisfaction or employee wellbeing, they struggle. Most organisations are measuring AI performance, not value.”
The distinction matters. Technical performance speaks to capability, but value speaks to consequence. For enterprises embracing AI at pace, the gap between the two is becoming increasingly visible.
Value before complexity
The concept of value-first AI reframes the conversation away from what systems can do toward why they exist in the first place. Odaro argues that organisations often adopt technology with an implicit assumption that sophistication automatically produces benefit. In practice, the opposite can occur, where highly capable systems deliver outcomes that customers neither notice nor trust.
“What we are really doing is building systems that encode our values into the future,” he says. “Model complexity tells us very little about whether we are solving problems that actually matter.”
The implications for executive leadership are significant. If value becomes the central metric, leaders must redefine how success is assessed across AI initiatives. Productivity gains alone do not provide a complete picture, particularly when employees feel disempowered or customers encounter opaque automated decisions they cannot challenge. This emphasis on value introduces a broader perspective on performance. Financial outcomes remain important, yet they sit alongside factors such as confidence, autonomy and long-term trust. These are harder to quantify but no less critical to sustainable adoption.
The shift also exposes a deeper tension in enterprise strategy. AI deployments are frequently justified through efficiency narratives, but efficiency does not necessarily equate to progress. When speed becomes the dominant objective, human experience risks becoming collateral damage.
Human centred means human agency
Human-centred AI has become a widely used phrase, but Odaro suggests that its meaning has been diluted. In many boardrooms it functions as a branding statement rather than a principle guiding difficult decisions. “Human centred has almost become a marketing term,” he says. “Real human centred design takes courage. It means being willing to say no to AI, even when it is profitable, if it erodes human agency.”
That willingness to decline certain uses of AI represents a radical departure from prevailing industry thinking. It demands that organisations ask whether technology makes people more capable or merely more dependent on algorithms. In this framing, the purpose of AI is not to replace judgement but to enhance it.
The challenge becomes particularly visible when discussing trust. Many organisations measure trust through adoption rates, assuming that widespread use implies confidence. Odaro rejects that assumption. “Adoption is often treated as trust, but that is theatre,” he says. “People use systems because they must, not because they believe in them. We need to ask whether users feel confident enough to disagree with the AI.”
He describes this as informed dissent, a concept that reframes trust as the ability to challenge recommendations without fear. Metrics that capture dissent, confidence and critical engagement may provide a more realistic picture of how humans and machines are interacting. For senior executives, this introduces a new set of key performance indicators. Safety, wellbeing and confidence become operational metrics rather than abstract ethical aspirations.
Culture before technology
Across industries, AI is still widely treated as a technology stack problem. New platforms are deployed, tools are integrated, and transformation programmes are launched under the assumption that technology will reshape behaviour naturally. Odaro argues that this sequence is fundamentally backwards. “Tech centres often believe that technology can solve culture,” he says. “But I have seen organisations deploying advanced AI while employees do not even want to share basic data with each other. The problem is not technical. AI adoption is a human transformation initiative.”
This perspective resonates with the growing recognition that behavioural change determines whether AI initiatives succeed. Systems built from data generated by people inevitably reflect existing organisational dynamics. If collaboration is weak or incentives misaligned, AI simply amplifies those issues at scale.
The emphasis on culture also alters the leadership conversation. Responsibility for AI cannot sit exclusively with technical teams. Values, incentives and behaviours must be shaped at the highest levels of the organisation. “AI is made from our data and by our decisions, but we forget that,” Odaro says. “We centre everything around the technology and ignore the humans behind it.”
For enterprises seeking competitive advantage through AI, this creates an uncomfortable reality. Transformation requires more than procurement and infrastructure investment. It requires confronting long-standing behavioural patterns that technology alone cannot fix.
Adaptive governance as a leadership function
Governance has long been perceived as a constraint on innovation, yet the speed of AI development is forcing organisations to reconsider that assumption. Traditional compliance frameworks struggle to keep pace with technological change, resulting in rigid policies that quickly become obsolete.
Odaro believes the governance conversation starts with the wrong question. “Executives ask how to control AI risk,” he says. “The better question is how to govern human decision-making when AI is involved. This subtle shift places accountability back where it belongs, with people rather than systems. Many regulatory frameworks focus heavily on technical compliance, but organisations often struggle to define who is ultimately responsible when outcomes go wrong.
“When I ask who owns the impact of AI on employee wellbeing or customers, most point to IT. That is like making the facilities manager responsible for company culture.”
Adaptive governance, in his view, relies on principles rather than rigid rules. Governance frameworks should evolve alongside technology while maintaining clear accountability for outcomes. Instead of prescribing every possible scenario, they must provide guidance that enables informed decisions in unfamiliar situations.
This approach requires executives to map existing decision-making processes and understand where AI intersects with them. The objective is not to slow innovation but to ensure that innovation aligns with organisational values.
Imagining futures before they arrive
Perhaps the most overlooked risk in AI development is not the known dangers that dominate public debate, but the blind spots created by limited imagination. Discussions often focus on bias, automation and job displacement, yet deeper societal consequences remain largely unexplored. “We know a lot less than there is to know,” Odaro says. “Tech leaders often assume efficiency equals societal progress, but we are building systems where the human side is diminishing. People feel that in everyday interactions.”
He warns that without deliberate effort; AI could gradually optimise for measurable outcomes at the expense of meaning. The danger is not a collapse, but a slow erosion of empathy, creativity and independent thinking. “Every major platform is embedding AI into everything,” he says. “The decisions made in boardrooms today will shape how future generations think and relate to each other.”
This framing shifts the responsibility beyond corporate performance metrics toward broader social stewardship. Organisations deploying AI are not only designing tools but also influencing patterns of behaviour at scale.
The stakes become clearer when considering the alternative. A future driven solely by efficiency could lead to increasing dependence on algorithmic guidance, reducing human agency in subtle but profound ways. “If AI makes people more dependent rather than more capable, that is not innovation,” Odaro says. “That is digital dependency.”
The challenge for executives is therefore not simply to innovate quickly, but to imagine the consequences of success. Blind spots emerge when leaders fail to ask what kind of society their technology is helping to create.
Leadership at the centre
Throughout the discussion, responsibility repeatedly returns to leadership. AI strategy cannot be delegated entirely to technical departments because the decisions involved are fundamentally moral and cultural. “AI should amplify existing organisational values,” Odaro adds. “If those values are unclear or purely profit-driven, AI will scale that problem. That is a chief executive issue, not a technical issue.”
The message is not anti-technology. Instead, it calls for a more deliberate alignment between innovation and human outcomes. Leaders must define what good looks like before technology defines it for them. In practical terms, this means expanding the definition of value. Financial performance remains essential, but it must sit alongside measures of trust, capability and societal impact. Organisations that can align these dimensions may find that AI strengthens rather than destabilises their culture.
As the pace of adoption accelerates, the window for shaping these choices narrows. The rush to deploy increasingly powerful systems leaves little time for reflection, yet reflection may be precisely what is required. “The future depends on whether we choose to encode values that enhance human decision-making,” Odaro continues. “We have one big chance to get this right.”
The question facing enterprises now is not whether AI will transform society. That transformation is already underway. The real question is whether leaders will have the imagination and courage to ensure that the systems they build create a future worth inheriting. In an era defined by rapid innovation, the most radical decision may be to slow down just enough to ask what progress truly means.
