The director’s role in a human-centred AI strategy
Boards do not need to become technical specialists, but they do need to shape how AI is governed, explained and used. The director’s role is to balance experimentation with accountability, ensuring AI serves people, supports decisions and stays aligned to strategy, risk and culture.
AI is moving quickly into everyday work, yet the real board question is not whether to adopt it, but how to keep it human-centred. Research from Stanford suggests capability is advancing fast while responsible AI practices are less mature, and incidents continue to be documented across the field 1. That gap matters for directors because strategy, culture and risk appetite are board responsibilities, even when the technology itself sits with management. Evidence from Harvard also suggests leadership quality remains visible when people work with AI, while MIT points to the need for clear governance and distinct approaches for broad versus tailored use 23.
What human-centred AI means for directors
A human-centred AI strategy is not a slogan about kindness. In practical board terms, it means AI is introduced to improve judgement, service and productivity without eroding trust, accountability or respect for people. That requires directors to ask a simple question: where will AI genuinely improve work, and where might it quietly weaken it?
The Stanford AI Index shows that adoption is widespread and capability is improving, but responsible AI reporting remains patchier than capability reporting 1. For directors, that suggests a familiar governance pattern: enthusiasm can outpace control. The board should therefore expect management to define the intended human benefit of each AI use case, not just the productivity gain.
An understated Authenticate advisory perspective is that directors often add the most value when they keep the conversation anchored to purpose: if AI is not improving a human decision, workflow or service in a defensible way, it probably needs a sharper case for adoption.
Separate broad use from tailored use
MIT Sloan distinguishes between broadly applicable generative AI tools and tailored solutions built for specific business needs 3. That distinction matters because the governance approach should change with the use case. Broad tools need clear acceptable-use rules, privacy safeguards and training. Tailored solutions need a more formal process for deciding whether to buy, build or adapt, together with closer oversight of vendors and dependencies.
Directors should ask management to map AI applications into these two categories and identify who owns each decision. Without that clarity, organisations can drift into shadow AI, where teams adopt tools informally and outpace governance. The practical board test is whether the organisation can explain, in plain language, what is allowed, what is not, and who is accountable when AI is used in work.
A human-centred strategy also recognises that control is not the same as resistance. Sanctioned access to a limited set of tools, backed by guidance and evaluation standards, is usually safer than pretending staff will not use AI at all 3. The board’s role is to insist that experimentation happens inside a defined frame.
Treat leadership as a human skill, not a technical badge
Harvard Business School’s discussion of leadership with AI agents suggests that effective leaders do not simply talk more; they ask more questions, encourage turn-taking and adjust when new information emerges 2. While the study is a laboratory experiment rather than a field trial, it points to a useful board implication: leading with AI still depends on listening, curiosity and disciplined judgement.
For directors, this matters in two places. First, in board deliberations, AI proposals should be interrogated with the same rigour as any major strategic investment. Second, in executive capability, leaders should be assessed on whether they can frame problems clearly, question assumptions and make room for dissent. If management cannot explain AI choices in a way that invites challenge, the strategy is probably too brittle.
Boards should therefore request evidence of leadership readiness alongside technical readiness. That might include scenario discussions, escalation pathways and regular review of how management is using AI to support decisions, rather than allowing automation to become a substitute for oversight.
Build governance around risk, data and accountability
The MIT research is clear that organisations need cross-functional guidance involving technology, legal, privacy and governance interests 3. For boards, that means AI oversight cannot sit in a single function. A human-centred strategy should specify who approves use cases, who monitors model behaviour, who reviews incidents and who can pause deployment if risk rises.
Directors should also insist on a plain-English register of AI use cases. It should distinguish between employee productivity tools, customer-facing tools and decision-support systems. The board does not need technical detail in every paper, but it does need enough information to understand the sensitivity of data, the level of human review, and the likely consequences if outputs are wrong, biased or misleading.
The Stanford report’s finding that safety and accuracy can pull in different directions is a reminder that trade-offs are real 1. Directors should ask management how those trade-offs are being weighed, documented and reviewed over time. If a team says an AI system is safe because it is efficient, that is not enough. Safety, usefulness and accountability all need to be examined together.
Keep people at the centre of adoption and change
Human-centred AI is also a change leadership issue. MIT notes that employees need trust, clarity and a sense of direction when AI enters the workplace 4. Directors should expect management to explain how staff will be prepared, how concerns will be addressed and how capability will be built across functions and career stages.
In practical terms, boards should ask whether the organisation has training that goes beyond prompting tips. People need to understand where AI can support work, where it must be checked, and when human judgement must prevail. They also need permission to raise concerns without penalty, especially if outputs appear plausible but are unreliable.
The strongest board contribution is to keep the conversation grounded in service of people and purpose. That means watching for overreach, avoiding vague promises and asking for evidence that AI is being used to strengthen, not thin out, the human elements of leadership, decision-making and culture.
Sources
- The 2026 AI Index ReportStanford University — Human-Centered AI · University research · link checked 15 October 2026
- Think You’re a Good Leader? Try Leading an AI TeamHarvard Business School AI Institute · University research · link checked 15 October 2026
- How to manage two types of generative artificial intelligenceMIT Sloan School of Management · University research · link checked 15 October 2026
- Who should be the head of generative AI — and what they should doMIT Sloan School of Management · University research · link checked 15 October 2026
Further professional reading
External professional resources, separate from the numbered sources above. These are not university studies, and listing them does not indicate any collaboration, certification, affiliation or endorsement.
- Global Institute of Directors (opens in a new window)External professional resource · governance and directorship
- World Coaching Institute (opens in a new window)External professional resource · coaching practice