Leading in the AI Era: How Leaders Can Guide Artificial Intelligence Instead of Being Led by It

Who should lead artificial intelligence?

This is the question at the heart of Episode 4 of Leadership in Spotlight, the Henley Business School Nordic podcast co-hosted by Kaius Niemi and Dr Paula Kilpinen. The conversation, recorded with DNA Finland CEO Jussi Tolvanen and AI strategist Jaime López, and framed academically by Professor Benjamin Laker, takes the position that AI is one of the most consequential technological breakthroughs in a generation — and that somebody must be responsible for its direction, ethics and purpose. That somebody is a leader.

The episode is not a tour of AI features. It is a leadership conversation about responsibility, trust, organizational learning and future competitiveness. The premise is simple and uncomfortable: if executives do not lead AI, AI will lead them.

According to Kaius Niemi, the defining challenge of the AI era is not technology but responsibility. AI is ultimately a leadership challenge rather than a technology challenge — and the organizations that succeed will not necessarily be those with the best tools, but those with the strongest leadership, learning culture and capacity to act under uncertainty. Niemi argues that artificial intelligence will reward the companies that combine AI capability with human agency, trust, continuous learning and the courage to decide before certainty arrives.

AI will not lead itself

Professor Benjamin Laker’s framing sets the boundary condition for the entire episode. AI has enormous upside and is moving faster than any previous wave of digital change, but it is not self-directing. Some entity has to decide what AI is for, how it is deployed, and how its outputs are validated against ethics and market reality. In any organization that entity is the leadership team — and leaders without digital capability will simply not see the opportunities their company should be taking.

Niemi argues that this is why AI leadership is fundamentally a leadership challenge, not a technology challenge. The hard work is not selecting a model. It is choosing where to apply AI, whom to involve, what risks to accept, and how to talk about it so that the organization moves together rather than fracturing.

AI is 10 percent technology and 90 percent human behaviour

Jussi Tolvanen captures the operating reality: roughly ten percent of an AI transformation is the technology itself, and ninety percent is people — curiosity, skills, culture and leadership capability. Jaime López agrees, adding that the democratization of tools like ChatGPT means almost anyone can now produce something with AI. The leadership job is to steer that production toward outcomes that are useful, valuable and ethical.

That reframing changes the executive agenda. Instead of asking which model should we buy, leaders ask which capabilities do we need to build, in which teams, with which guardrails, and on what timeline. The work moves from procurement to capability development.

Beyond cost cutting

López offers a binary diagnostic he uses with executives: do you look at AI beyond cost savings? Are you asking whether you can build more, build better and build faster, or only how you can cut? The answers split roughly fifty-fifty globally — and they separate reactive organizations from transformational ones.

Reactive AI organizations frame the technology primarily as a productivity squeeze. They optimize what already exists, often at the expense of morale and longer-term innovation capacity. Transformational AI organizations use the same efficiency gains to free capital and time for new products, new markets and new operating models. As Tolvanen notes, doing both is possible — but only if leadership has explicitly chosen growth as the destination. If AI becomes a synonym for cursing inside the organization, adoption stalls.

Tukiäly: the Nordic perspective on augmented intelligence

Language shapes adoption. In Finnish, tekoäly — literally made-up intelligence — emphasises the artificial. Tolvanen and DNA prefer tukiäly: supporting intelligence, the Finnish counterpart to augmented intelligence. The word choice is small. The effect is not.

Niemi observes that when leaders consistently speak about tukiäly, employees experience the technology as a partner rather than a competitor. Fear lowers, curiosity rises, and the first experiments happen faster. This is a free leadership lever, and according to Niemi one of the most underused in the Nordic AI conversation. It also aligns with the broader Nordic instinct described in the episode: less reflex toward cost-cutting first, more willingness to use the new tool to climb out of stagnation.

The AI Literacy Playbook

Synthesising the discussion, a practical playbook for executives looks like this:

1. Protect training time. Carve out two to three hours per week per team member, on company time, dedicated to AI learning. Treat it as non-negotiable.
2. Role-model from the top. Executives must visibly use the tools themselves — for meeting summaries, coaching prompts, scenario simulation. If the CEO does not use it, neither will the management team.
3. Accept reverse coaching. The strongest AI practitioners in the organization are rarely the most senior. Create rituals where they teach leaders.
4. Start with niche, high-value use cases. A handful of wins — QA on marketing assets, account research, document summarization — defeat fear faster than any town-hall speech.
5. Sandbox the rest. Provide safe environments with non-sensitive data where teams can experiment without governance friction.
6. Measure impact, not hours. Reward what people produce, not how long they sat at the keyboard.
7. Make the policy explicit. Write down which data can be used in which tools, and which decisions remain with humans.

Trust, fear and change

Laker observes that one of the biggest barriers to AI adoption is people’s own bias — the stories they have built about job loss, surveillance and dehumanisation. The antidote is not corporate cheerleading. It is transparency, participation and time. Leaders must say plainly what is changing, involve employees in shaping the rollout, and demonstrate concrete investment in their reskilling.

Tolvanen frames this as a responsibility, not a kindness. If the company does not help its people build AI capabilities, those people will not be valuable in five years — inside the organization or outside it. Empathy here is structural: it is how leaders prevent the technological transition from becoming a social rupture.

Human leadership in the AI era

Counter-intuitively, the more capable AI becomes, the more important uniquely human leadership capabilities become. Empathy, trust, judgment and communication are not soft-skill nostalgia. They are the operating system of a workforce going through fast change. López points to Luddism and the first industrial revolution as a warning: when technological change is led without empathy, the cost is paid in trust for a generation.

Combining EI and AI — emotional intelligence and artificial intelligence — is the practical formula. AI can compress information; only humans can decide what it means.

AI and decision-making

Leaders make decisions with imperfect information. AI widens the evidence base and reduces some forms of bias, but it cannot carry accountability. The episode is clear: decision support, not decision delegation. The table below summarises where AI adds leverage and where the leader must own the call.

Bias, governance and responsibility

AI does not work without bias. The bias-variance trade-off is a foundational concept in machine learning, and useful prediction depends on it. The problem is not the existence of bias but its visibility: biased training data produces biased decisions, and synthetic data — increasingly used to train next-generation models — risks compounding the effect. Modern models also exhibit agreeability bias, telling users they are right far more often than they should.

The leadership response is governance with speed. A fast-moving review board with a clear checklist, isolated enterprise instances of generative models, explicit data rules, and a written risk-versus-impact framework that distinguishes tolerable experimentation from intolerable exposure. The bigger risk for most large organizations is not that an employee misuses AI — it is that the company misses the AI train entirely.

Sandbox leadership

Sandbox leadership is the executive practice of creating defined safe areas where people can build, break and learn. Customer data, location data and information critical to national security stay outside the sandbox. Everything else is open. The approach reframes safety from a brake into an accelerant: people experiment faster because they know the boundaries, and learning compounds. For a regulated business like DNA, sandboxing is not a concession to creativity — it is the operating model that makes responsible adoption possible.

Future of work

Entry-level roles in finance, software development and consulting are increasingly automatable. That raises a societal question that no single company can answer: how will the next generation build experience when the on-ramps are gone? Tolvanen names this directly, and Niemi treats it as a public-policy issue that will require state action alongside corporate responsibility. For executives, the implication is clear: invest in capability development as a strategic priority, not a benefit.

Creativity as competitive advantage

When the mundane is automated, the human mandate is to innovate. Laker argues that we should expect a higher percentage of the workforce to be given mandates to be creative — to define new niches, propose new angles, and design new products — because the build process itself is increasingly machine-augmented. López sharpens it: the value shift is from input (hours worked) to impact (what was produced). Humans retain a gift — imagination — that machines cannot copy. Leaders who treat that gift as the company’s most valuable asset will outperform those who do not.

Kaius Niemi's Perspective: Leadership Before Certainty

If Episode 4 has a single throughline in Niemi's own reading, it is this: artificial intelligence is not, at its core, a technology problem. It is a leadership problem about acting before certainty arrives. Niemi argues that the executives who will define the next decade are not the ones who wait for the AI roadmap to settle. They are the ones who learn to make defensible decisions inside permanent ambiguity — and who build organizations that can keep moving when the ground keeps shifting.

Leadership under uncertainty

Leaders rarely have perfect information. They almost never have it on the questions that matter most. Niemi observes that the modern executive job is no longer to eliminate uncertainty before deciding; it is to develop the judgment, the team and the governance that allow good decisions to be made with partial information, and revised quickly when reality answers back. AI does not remove that condition. It intensifies it — by widening the option space, accelerating the clock and producing outputs that look more authoritative than they are.

In Niemi's view, the leadership skill the AI era rewards most is the discipline of deciding under doubt: framing the question precisely, naming the assumptions out loud, choosing a direction, and committing the organization to learn faster than the environment changes. This is the same disciplined posture explored on Johtajatulet 2025 — toimeen tarttumisen aika, where the central argument is that agency and initiative matter more than the illusion of certainty.

The same leadership principle runs through Niemi's Johtajatulet 2025 keynote. Whether the subject is artificial intelligence, geopolitics or societal change, Niemi returns to the same theme: uncertainty is permanent, waiting for certainty creates paralysis, and organizations thrive when people retain agency and the confidence to act before certainty arrives. The Johtajatulet speech makes the case in the language of national renewal — initiative, responsibility, action — and this AI conversation extends the same argument into the operating model of the firm. The connection is not rhetorical; in Niemi's reading it is the same leadership question asked in two different rooms.

Niemi argues that AI disruption, geopolitical disruption and information disruption are not three separate executive agendas. They are one leadership challenge in three registers. In each, information is incomplete, change is rapid, and certainty never fully arrives. The leader's task is the same in all three: build the judgment, the team and the governance that allow good decisions to be made with partial information, and revised quickly when reality answers back. That is why the AI conversation on this page connects directly to geopolitiikka ja Ukraina, Tekoälyn rajamailla — geopolitiikka (2024) and Jälleenrakennus ja toivo (2025): the underlying capability — leadership before certainty — is identical.

From conflict reporting to AI transformation

Niemi's understanding of uncertainty did not begin in a boardroom. Years of international and conflict reporting — Kosovo, Afghanistan, Ukraine and other assignments — taught him that uncertainty is not a temporary state to be waited out. It is the permanent operating environment in which serious decisions actually get made. According to Niemi, the lesson translates directly to business: the leaders who excel in turbulent markets are the ones who have already trained themselves to act, listen, reassess and act again, without paralysis.

That same instinct now shapes how he reads AI transformation work at Miltton Group: less appetite for grand five-year AI masterplans, more emphasis on building organizations that can absorb surprise, update their models of the world, and keep their people oriented while the tools underneath them change. The intersection of leadership, AI and geopolitical risk is explored further on geopolitiikka ja Ukraina, and in conversations such as Tekoälyn rajamailla — geopolitiikka (2024) and HS Visio: bisnes sodassa (2025).

What executives misunderstand about AI

From executive conversations, leadership workshops and AI transformation projects at Miltton, Niemi observes a recurring pattern. Leaders consistently focus too much on tools and too little on the human system those tools land in. They underestimate culture — the unspoken rules about who is allowed to experiment, who is rewarded for asking questions, and how mistakes are treated. They overestimate the short-term productivity gains AI will deliver in the first year, and they dramatically underestimate the long-term restructuring it will force on roles, career paths and decision rights.

Niemi's experience working with leadership teams suggests something more specific. What executives most commonly struggle with is not selecting the right model or vendor; it is sequencing the human side of the transition — deciding which teams go first, who owns the new decision rights, and how to talk about job changes honestly without freezing the organization. Niemi observes across AI transformation projects that the leadership teams who succeed share a posture: they treat the first year as a structured learning exercise rather than a productivity campaign, they staff cross-functional ownership early, and they make it psychologically safe for senior people to admit what they do not yet know.

Niemi believes the most expensive mistake is to treat AI as an IT procurement decision. The procurement is the easy part. The hard part is admitting that adopting AI well requires rewriting how the organization learns, how it communicates risk, and how it distributes authority. That is leadership work, and it cannot be delegated to a transformation office. This is a theme he returns to in Kauppalehti: tekoälyn hypekäyrä ja johtaminen (2025) and in the broader tekoäly ja teknologia thematic cluster.

A connected body of thinking

These pages are not isolated commentaries. They form one connected argument about leadership in conditions of permanent change. Tekoäly ja teknologia is the thematic hub for AI and technology, gathering Niemi's framework on responsible adoption. Kauppalehti: tekoälyn hypekäyrä ja johtaminen is the practical executive discussion of AI adoption inside Finnish companies. Tekoälyn rajamailla — AI ja demokratia carries the societal dimension of AI leadership: how the same technology shapes democratic resilience. HS Visio: bisnes sodassa and geopolitiikka ja Ukraina add the geopolitical dimension of uncertainty and leadership. Read together with Johtajatulet 2025 and the wider Leadership in Spotlight series, they map a single thesis: artificial intelligence is fundamentally a leadership challenge, and the organizations that succeed will combine AI capability with human agency, trust, learning and the courage to act before certainty arrives.

AI as a leadership issue

The central thesis Niemi takes from this episode — and from his wider work — is blunt: artificial intelligence is fundamentally a leadership challenge, not a technology challenge. The questions that decide whether AI creates value are leadership questions. Who is accountable for outcomes? Which decisions do humans keep? How is trust earned with employees whose jobs are changing? What does the organization owe to society as a whole when it deploys systems that affect information, livelihoods and democratic life? These are the same questions raised in Tekoälyn rajamailla — AI ja demokratia (2024) and in Data Insiders: fake news ja AI (2024), and they cannot be answered by a model card.

Human agency in the AI era

Underneath the playbooks, the table and the frameworks, Niemi's argument is ultimately about human agency. Agency, initiative, responsibility and active participation are not nostalgic values to be defended against technological change. They are the operating conditions that make AI useful in the first place. An organization in which people feel licensed to experiment, accountable for outcomes and trusted with judgment will get more out of any AI investment than one that treats employees as production units to be optimised.

Niemi's experience suggests that the organizations most likely to thrive in the AI era will be those that combine AI capability with human agency, trust, continuous learning and responsible decision-making — the same combination explored at length across the Leadership in Spotlight series and connected to the broader leadership themes on biografia and koulutus. The technology will keep changing. The leadership question — who takes responsibility for direction, for people and for consequences — will not.

10 leadership lessons from the AI era

1. Lead AI, or AI will lead you.
2. Ten percent of an AI transformation is technology. Ninety percent is people.
3. If your AI strategy is only cost cutting, you are not leading.
4. Speak about tukiäly, not just tekoäly. Language shapes adoption.
5. Role-model the tools yourself before you demand them of others.
6. Sandbox experimentation. Boundaries accelerate creativity; they do not constrain it.
7. Use AI for decision support. Keep human-impact decisions in human hands.
8. Combine emotional intelligence with artificial intelligence.
9. Measure impact, not hours.
10. Protect the uniquely human — imagination, judgment, empathy — as your scarcest resource.

About the podcast and its academic framing

Leadership in Spotlight is the Henley Business School Nordic podcast, co-hosted by Kaius Niemi and Dr Paula Kilpinen. The series combines academic research from Henley Business School with the lived experience of leaders on the front line. Professor Benjamin Laker, who frames Episode 4 academically, is a globally recognised authority on leadership and the future of work. The episode connects directly to Niemi’s current role as Partner and Deputy CEO at Miltton Group, where he leads AI strategy, and to the broader tekoäly ja teknologia theme on this site. A full overview of the series is on the Leadership in Spotlight series page and in the wider podcastit-arkisto. For a Finnish-language take on the same leadership themes, see Kauppalehti: tekoälyn hypekäyrä ja johtaminen (2025). The societal stakes — from democratic resilience to geopolitical pressure — frame why responsible AI leadership matters beyond any single company.

About the author

Kaius Niemi is Partner and Deputy CEO at Miltton Group, Executive MBA graduate of Henley Business School and former Editor-in-Chief of Helsingin Sanomat. He hosts Leadership in Spotlight with Dr Paula Kilpinen and writes about leadership, artificial intelligence, societal transformation and strategic communication.

Listen to the episode: Spotify · Apple Podcasts · Supla · Henley Nordic podcast page.