Artificial intelligence can help us do work that once seemed beyond our reach. It can also make a mistake look convincing, multiply deception and put powerful capabilities into irresponsible hands. Both possibilities belong in the same conversation.
I regard AI as one of humanity’s greatest achievements since the printing press. We should keep developing it and exploring what it can do. That conviction comes with a responsibility: to establish standards, safeguards and accountability that match its growing power.
A wider field of possibility
The printing press transformed how knowledge circulated. Over time, it enabled ideas to reach far beyond the people and institutions that had previously controlled their distribution.
AI could prove similarly consequential. It lowers barriers to activities that previously demanded substantial time, money or specialist support: developing software prototypes, exploring complex information, working across languages and testing new ideas.
Scientific research offers a striking example. AlphaFold’s breakthrough in predicting protein structures has given researchers a powerful resource for investigating biological questions. Its predictions still require careful interpretation. The achievement shows how AI can extend human inquiry while leaving essential work for scientists to do.
For smaller businesses, the opportunity is equally worth exploring. A modest team can investigate options and develop early prototypes that it might previously have struggled to resource. Access to those capabilities creates room for ambition. Turning them into dependable results still takes expertise.
More output is only the beginning.
Much of the business conversation focuses on speed: how quickly AI can write a report, produce an image or generate code.
Time saved can be valuable. But a report produced in seconds may contain an unsupported conclusion. A polished proposal may promise something the business cannot deliver. More content can consume attention without helping anyone make a better decision.
Leaders therefore need to ask what the additional capability is for.
Consider a small manufacturer exploring a new service. AI could help the team organise customer feedback, identify recurring problems and create an early prototype. Staff would check the interpretation against actual customer conversations, test whether the service solves a worthwhile problem and assess whether the business can deliver it.
The gain comes from investigating an opportunity more thoroughly with the resources available. The people who select the evidence, evaluate the proposal, and decide what happens next own the responsibility.
Responsibility must be visible.
“Use AI responsibly” sounds reassuring. It becomes meaningful when someone can explain what that requires in practice.
Who checks an output before a customer relies on it? What information may staff share with a system? Who can stop an automated process when it behaves unexpectedly? Who investigates and corrects harm?
These questions belong in the work's design from the beginning. A person assigned to review an AI recommendation needs the knowledge, time and authority to challenge it. A signature at the end of a process provides little protection if the reviewer cannot assess what happened.
The level of scrutiny should reflect the consequences. Drafting an internal meeting agenda and influencing someone’s access to an essential service demand different controls.
Developers, providers, business leaders and users each carry responsibilities. Those responsibilities need to be explicit enough that accountability survives when something goes wrong.
Optimism requires clear boundaries.
My support for AI’s development includes support for safeguards against misuse.
Fraud, impersonation and manipulation undermine the trust on which useful adoption depends. In warfare, delegating decisions involving human life raises especially serious questions about control and accountability.
We should confront these issues directly while continuing to pursue beneficial applications. Clear standards can help people adopt AI with greater confidence, understand its limits and challenge harmful uses.
Leaders cannot determine the technology’s entire trajectory. They can choose what they deploy, whose interests they consider, what evidence they require and where they retain human authority.
The questions ahead
This five-part series will explore how to use AI effectively while preserving judgement, protecting trust, addressing misuse and setting boundaries in high-stakes situations. It will also examine how businesses can turn new capabilities into worthwhile services and better ways of working.
For your leadership team, start with two questions: What could AI enable us to do that would create real value? And who will take responsibility for how we do it?
AI expands our capabilities. Responsibility remains ours.
Next: When AI Gives You an Answer, Who Does the Thinking?
MaxMORIX
EXPERTS helps leadership teams explore where AI can create value and what needs to change in their business design. Discuss your situation with us.
