By Dr. Kevin Shepherdson
In my previous article, I discussed the case “Judge Learns Lawyers on Both Sides of Case Used AI, Cancels Trial, Kicks Everyone Off the Case.” I used it in my Gen AI courses as a vivid example of a broader pattern I see in HR and Learning & Development, where both sides increasingly rely on AI tools — teachers and students, hiring managers and candidates, each augmenting their work with AI. This pattern is not unique to law. It is spreading across professions, institutions and even personal relationships, from software development and insurance claims to university admissions and online dating. In each of these domains, AI is now present on both sides of the table, quietly reshaping how work is done, how people are evaluated and how trust is formed.
The real problem: passive AI use
Let me be clear: I am not against AI. AI can be an extraordinary tool.
It can enhance productivity, improve learning, support decision-making, widen access to expertise and help people create things they could not previously create.
The problem is not AI use, but passive AI use. Passive AI use happens when people accept AI outputs without questioning them — when professionals copy, paste, submit, approve, publish or decide without understanding the reasoning behind the output.
It happens when AI becomes a shortcut around thinking rather than a tool for thinking better. The danger is not that machines will replace us overnight. It is that humans will slowly withdraw from judgment.
From human-in-the-loop to human-on-the-hook
I often hear GRC professionals emphasise the importance of having a “human in the loop” in AI. But this phrase can be misleading.
A human in the loop is not useful if the human does not understand the task, cannot challenge the output, lacks time to verify the result or feels pressured to approve what the AI has produced. In such cases, the human is not truly in the loop. The human is merely in the workflow.
Worse, the human may become a rubber stamp.
What we need is not just human-in-the-loop. We need human-on-the-hook.
That means someone remains accountable. Someone verifies. Someone challenges. Someone understands the limits of the AI system. Someone owns the final decision.
1. In law, the lawyer remains accountable.
2. In education, the teacher and student remain accountable.
3. In HR, the hiring manager remains accountable.
4. In software, the developer and reviewer remain accountable.
5. In governance, the organisation remains accountable.
AI may assist the work. But accountability cannot be outsourced to the machine.
The illusion of productivity
One of the biggest promises of AI — which I often share in our workshops — is that it can enhance productivity, streamline operations and increase the value of products and services. And yes, AI can produce more output, faster.
But we must be careful not to confuse output with value.
1. A longer report is not necessarily a better report.
2. A faster legal submission is not necessarily a sound legal argument.
3. A polished CV is not necessarily a stronger candidate.
4. A fluent essay is not necessarily evidence of learning.
5. A working code snippet is not necessarily secure software.
AI makes it easier to produce. But organisations must ask: produce what, for whom, with what quality and under whose accountability?
If we do not ask these questions, AI may create an illusion of productivity — more documents, more messages, more analysis, more dashboards, more content — while weakening the very judgment needed to use those outputs well.
The new digital arms race
When both sides use AI, we also create a new digital arms race.
1. Students use AI, so schools use AI detectors.
2. Applicants use AI, so employers use AI screeners.
3. Marketers use AI, so consumers use AI filters.
4. Hackers use AI, so defenders use AI security tools.
5. Claimants use AI, so insurers use AI fraud detection.
6. Lawyers use AI, so courts demand AI disclosure and verification guidelines.
Each side escalates. The result is not always better work. Sometimes it is simply more automation, more suspicion and more complexity.
This is why organisations must move beyond the shallow question of “Which AI tool should we use?”
The better question is: what human capability must we strengthen so that AI does not weaken our judgment?
AI capability is not the same as AI usage
This is a distinction we often make in our work at Straits Interactive. Using AI is not the same as building AI capability.
Usage is when people use tools.
Capability is when organisations combine skills, knowledge, tools, processes, governance and transformation to create sustainable value. This distinction matters.
1. A student using AI has usage. A school redesigning assessment, teaching critical thinking and setting clear AI rules is building capability.
2. An HR manager using AI to screen CVs has usage. An organisation redesigning recruitment workflows with fairness checks, human oversight, explainability and accountability is building capability.
3. A lawyer using AI to draft submissions has usage. A law firm implementing verification protocols, professional standards, review processes and AI governance is building capability.
The winners in the AI era will not simply be those who use AI the most. The winners will be those who know when to use AI, how to challenge it, how to govern it and how to retain human judgment where it matters most.
The question leaders must ask
For business leaders, educators, HR professionals, lawyers, compliance officers and board members, the central question is no longer:
Are our people using AI?” They probably are.
The better questions are:
1. Are they using it well?
2. Are they checking the outputs?
3. Are they preserving critical thinking?
4. Are they clear about accountability?
5. Are they redesigning workflows responsibly?
6. Are they documenting decisions?
7. Are they aware of the risks?
8. Are they still learning, or merely producing?
AI transformation is moving quickly. But speed without judgment is not transformation. It is acceleration without steering.
Are we controlling the machines?
This brings me back to the legal article that started this reflection. When lawyers on both sides rely on AI carelessly, the issue is not merely professional embarrassment. It is a warning about what happens when institutions built on human judgment begin to outsource that judgment without adequate safeguards.
The same warning applies across education, HR, research, finance, insurance, software and even relationships. AI is not just changing how work is done. It is changing how trust is formed, how evidence is assessed, how decisions are made and how humans relate to one another.
So we must ask:
Are machines helping us think better?
Or are they quietly thinking on our behalf?
Are we using AI as a tool?
Or are we becoming the tool through which AI-generated outputs move from one system to another?
The future of AI should not be a world where machines talk to machines while humans merely approve, forward, submit and comply. The future of AI must be one where humans remain active, critical, accountable and capable.
Because when AI is on both sides of the table, the most important question is not which side has the better machine.
The most important question is: who is still doing the thinking?
This is Part 2 of a two-part series. Read Part 1: Who Is Doing the Actual Work? When AI Is on Both Sides of the Table here.
