AI Won’t lead itself
Mastering AI by Jeremy Kahn was an excellent book I read on the subject. His main premise is that the light-bulb moment for AI came on November 30, 2022, with the launch of ChatGPT. What distinguishes this wave of technology is the extraordinary pace at which its capabilities are advancing and spreading across the economy.
For me, that moment represents a turning point in public awareness—not the beginning of AI itself. The leadership challenge is to translate that awareness into responsible action, without confusing the speed of adoption with the creation of lasting value.
One of Kahn’s concluding ideas particularly resonated with me: mastering AI ultimately requires us to master ourselves—to bring our own intelligence, creativity, wisdom, and courage to how we develop and use this technology.
I agree. And that is what makes this not simply a technology challenge, but a leadership challenge.mThe question for leaders isn’t whether AI will change our organizations. It is how we guide that change.
My perspective is that adoption alone is not the objective. Better outcomes for customers, employees, and the business should be the measure of success.
Uses of Generative AI and the Industries It Is Impacting
Generative AI has rapidly moved from experimentation into real-world applications across industries. Increasingly, the opportunity is not simply generating content but augmenting human capabilities, improving decisions, redesigning workflows, and creating new ways of delivering value.
Here are several areas where that change is already visible:
Content Creation: Modern multimodal AI can work across text, audio, images, video, and code, enabling marketers, writers, designers, and developers to accelerate workflows and create increasingly adaptive content. Capabilities and reliability vary by model and task. Creative judgment, factual verification, and brand accountability must remain part of the workflow.
Marketing and Advertising: Generative AI can help marketers develop campaigns, personalize content, analyze customer behavior, and create experiences tailored to different audiences. The opportunity goes beyond producing more content—it is using intelligence to make engagement more relevant.
Healthcare: AI is being applied to areas including synthetic medical data, drug discovery, clinical research, diagnostics, and personalized medicine. The potential is enormous, but these opportunities should not be confused with uniformly proven clinical outcomes. Applications influencing patient care require rigorous validation, professional oversight, and appropriate regulatory review.
Entertainment: AI is enabling new approaches to virtual characters, music composition, scriptwriting, visual effects, gaming, and immersive experiences. It can accelerate production while opening new possibilities for personalization and creativity.
Finance: AI is being used across fraud detection, risk assessment, financial analysis, forecasting, and decision support. An important distinction is that many of these applications combine generative AI with predictive AI, machine learning, and traditional analytics. Generative AI should therefore be evaluated as part of a larger analytical system rather than treated as a guarantee of better forecasts or financial decisions.
Manufacturing: AI can support product design, supply-chain optimization, simulation, predictive maintenance, and production planning. These solutions often combine generative AI with optimization, simulation, computer vision, and other technologies. Ultimately, the benefits need to be demonstrated in the operating environment rather than assumed.
Across all of these industries, I would ask three questions before scaling:
What problem are we solving? How will we measure improvement? Who remains accountable when the system is wrong?
Those questions become even more important as AI systems become capable of taking action rather than simply generating recommendations.
Concerns of Generative AI
The opportunity is significant, but so are the risks. Responsible adoption requires leaders to confront them directly.
Ethics and Governance: The ability to generate realistic content raises concerns about deepfakes, misinformation, intellectual property, bias, and misuse. Enterprise governance addressing data sourcing, copyright, permissions, traceability, and audit trails is becoming foundational. Leaders need to determine what evidence must be retained, where human approval is required, and when an AI system should not be used at all.
Hallucination and Reliability: Generative AI can produce impressive results while still generating incorrect, incomplete, or misleading information. Approaches such as Retrieval-Augmented Generation can help ground responses in approved enterprise information, but retrieval does not eliminate errors. Access controls, source verification, testing, and output validation remain essential.
Data Privacy: AI systems can interact with enormous volumes of personal, proprietary, and sensitive information. This is particularly consequential in regulated industries such as healthcare and financial services. Protecting that information and complying with applicable regulations must be built into the architecture rather than added later.
Job Displacement: Automation will change jobs and could eliminate some tasks and roles while creating others. For me, the leadership objective should be to use AI wherever possible to augment human capability rather than simply replace it. Workforce development should therefore be part of the investment from the beginning—not an afterthought once implementation is complete.
The common thread across these risks is accountability.
AI may generate the recommendation. Increasingly, it may even execute the action. But responsibility for the consequences must remain with the people and organizations deploying the technology.
The Future of AI
Looking toward 2035, I believe the implications could become considerably more profound. These are scenarios rather than predictions, but they illustrate where today's capabilities could lead.
Human-AI Collaboration and Agentic Execution: AI will increasingly move beyond chat interfaces toward agentic workflows—systems capable of planning, executing multi-step operations, using tools, coordinating tasks, and working alongside people in creative, analytical, and operational processes. Agentic capabilities already exist. The longer-term opportunity lies in improving their reliability, governance, and integration at enterprise scale. If AGI is achieved, that evolution could accelerate, although neither its achievement nor timing is certain.
Healthcare and Disease: AI could accelerate drug discovery, improve prevention, support earlier diagnosis, enable more personalized treatments, and expand the capabilities available to clinicians and researchers. Eradicating diseases is a far higher bar and will depend on far more than technological progress, including clinical trials, regulation, healthcare infrastructure, economics, and access. But even partial progress could have enormous human impact.
Cybersecurity: AI will strengthen defenders' ability to identify patterns, detect threats, automate responses, and analyze vulnerabilities. At the same time, attackers will have access to increasingly sophisticated AI capabilities of their own. Leaders should therefore prepare for both more capable defenses and more sophisticated threats. Neither security nor compliance can be guaranteed by AI alone.
Entertainment: We are likely to see increasingly interactive, personalized, and immersive experiences that blur traditional boundaries between creator and audience and between physical and virtual environments. AI-generated content could become dynamically tailored to individual preferences.
The practical implication for leaders is not to predict 2035 perfectly. It is to build organizations capable of learning, adapting, and changing direction as the evidence develops.
In highly regulated industries such as insurance, healthcare, and financial services, the challenge is not simply deploying AI. It is deploying AI in ways that improve decisions while preserving trust, transparency, compliance, and accountability. Technology may create capability, but trust determines whether that capability translates into lasting value. As AI becomes embedded in critical business processes, trust may prove to be the ultimate competitive advantage. In fact, the organizations that outperform in 2035 may not be the ones with the most advanced AI. They may be the ones that learn fastest from it.
What Leaders Should Do Now
Given the pace of change, waiting is itself a decision. Here is where I would start:
1. Pilot with a human in the loop. Start with a clearly defined business problem and establish a baseline. Measure quality, cost, speed, customer or employee impact, and risk. As organizations experiment with agentic tools, keep appropriate human review around consequential decisions and actions. Give reviewers the expertise, time, and authority to challenge the technology—not simply approve what it produces.
2. Set governance early. Define what data can be used, who is accountable, and how decisions and actions can be audited. For systems capable of taking action, specify permissions, approval thresholds, escalation paths, and conditions for stopping execution. Accountability should be clear before deployment, not after something goes wrong.
3. Invest in reskilling. Help people learn how to work with AI so that it expands what they are capable of accomplishing. Make learning relevant to actual roles and workflows. Involve employees in redesigning how work gets done, and measure whether AI improves the quality of their work—not only how quickly they complete it.
4. Be honest about uncertainty. Leaders do not need to pretend they know exactly where AI is heading. Tell your teams what you know, what you don't, and how you will adjust as the technology develops. Establish review points. Scale what works, redesign what falls short, and be willing to stop initiatives that do not justify their cost or risk.
For me, the leadership standard is straightforward:
Pursue measurable value. Protect trust. Bring people along.
Generative AI has extraordinary potential, but technological capability by itself does not determine whether that potential becomes progress. That will depend on the choices organizations and leaders make about where AI is applied, how it is governed, and how its benefits are shared.
That brings me back to the idea from Kahn's book that stayed with me: mastering AI starts with mastering ourselves. As AI becomes more capable, human judgment does not become less important. It becomes more consequential.
Our responsibility as leaders is not simply to embrace the technology or resist it. It is to understand it, challenge it, govern it, and apply it where it can create meaningful value while protecting the trust of the people we serve.
That is the job of leadership in turbulent times.
The test is not simply whether we can deploy more intelligent technology. It is whether we exercise the judgment, courage, and responsibility to put it to worthwhile use.
The views expressed here are my own and do not represent my organization.
Comments
Post a Comment