The debate around artificial intelligence in enterprise strategy is often trapped in a false binary: machine autonomy or human control. Either we allow AI systems to operate independently, or we insist that people remain firmly “in the loop”. In my experience, that framing is already becoming obsolete.
As AI becomes more autonomous, the real question for business leaders is where human judgment matters most. What should machines be allowed to do independently? What should trigger human intervention? And who remains accountable for the outcome?
This is what I mean by human-orchestrated AI: machines operating at speed and scale, while people set the objectives, boundaries and accountability within which they operate. For emerging markets across the Middle East and Africa, getting this balance right matters enormously.
AI has the potential to address constraints around skills, service delivery, financial inclusion, productivity and access to expertise. We should not slow adoption simply because the technology introduces risk. However, deploying AI first and asking governance questions later creates risks of its own; the better approach is risk-based.
Low-risk, routine and reversible decisions can support greater automation. Decisions involving significant financial, employment, healthcare, safety, legal or societal consequences require much stronger human controls.
Think about the difference between an AI system categorising an internal document and one influencing whether somebody receives credit, gets shortlisted for a job or receives a healthcare recommendation. Applying the same oversight model to both makes little sense.
At scale, requiring a person to approve every AI output creates bottlenecks and undermines much of the value of automation. The human role is increasingly to design the loop – to establish the thresholds that determine when AI can act independently and when human judgment must take over. For emerging markets, however, human orchestration has another critical role: providing context.
AI systems are shaped by the data and assumptions on which they are built. Yet the languages, cultures, informal economies, socioeconomic conditions and customer behaviors found across emerging markets may not be adequately represented in global datasets.
The quality of the underlying data matters just as much. Human oversight cannot compensate for poor information. Organizations need to understand what data their AI systems use, whether it is representative, whether they have the right to use it and whether it introduces bias. In emerging markets, where the quality and availability of data can vary considerably, these questions need to be addressed from the outset.
Anyone who has worked across markets in our region knows that apparently sensible global assumptions can look very different when they encounter local reality. That is not an argument against global AI, rather one for combining global capability with local intelligence.
When AI influences decisions about credit, insurance, recruitment, healthcare or access to public services, local experts need to be able to recognize when a system has missed a linguistic or cultural nuance, misunderstood customer behaviour or relied on assumptions that do not translate well across markets.
Emerging markets therefore should not simply inherit oversight models developed elsewhere. We have an opportunity to apply global responsible AI principles in ways that reflect our own economic, cultural and regulatory realities.
There is one principle, however, that should apply everywhere: accountability cannot be delegated to an algorithm. As AI becomes more autonomous, organisations need clear answers to basic questions. Who authorised the system to act? Who set its boundaries? Who monitors it? Who can stop it? And who is responsible if something goes wrong?
This is why I do not see responsible AI as a brake on innovation. In fact, good governance can do the opposite. Clear expectations around accountability, fairness, privacy, transparency and integrity give organisations greater confidence about the boundaries within which they can innovate.
Emerging markets do not have to choose between rapid AI adoption and responsible AI adoption. We need to become more deliberate about where machines should have autonomy, where human judgment remains essential and how accountability connects the two.
