AI Changes What We Can Produce. Judgment Determines What We Should. (TPL Insights #325)
- Aug 13
- 5 min read
By Rob Andrews
A group of first-year marketing students recently presented an influencer campaign that looked remarkably polished. The influencer was articulate, attractive, completely on brand and entirely fake.
Artificial intelligence had created the avatar, the voice and much of the content.
The students used tools including HeyGen, Mango AI and ElevenLabs to build a 60-second influencer campaign around an existing brand. The technical execution was impressive. But that wasn't the most interesting part.
What happened next was.
Their classmates began asking questions.
Would consumers trust an influencer who didn't exist? Should the brand disclose that the person was artificial? What happens to authenticity when we can manufacture it? Could a synthetic spokesperson actually be more effective than a human one? And if it could, should that matter?
Suddenly, nobody was talking much about the software.
They were talking about judgment.
I have spent more than 30 years assessing executives and studying why some organizations consistently outperform others. More recently, I've spent a good deal of time thinking about what higher education does well, where it falls short and how we prepare people to lead organizations we can't yet imagine.
That classroom discussion captured something I think we're missing in much of the debate about artificial intelligence.
AI changes what people can produce. It raises the premium on judgment.
We Have Plenty of Answers
For most of my lifetime, education has operated under an understandable assumption. Knowledge is scarce, and educated people are distinguished in large part by how much of it they acquire and how effectively they can apply it.
That assumption is changing quickly.
A student can now ask an AI system to explain a theory, analyze a case, summarize research, write computer code, create an advertising campaign or produce a reasonably convincing argument in seconds.
That creates obvious academic integrity problems. We should take them seriously.
But I think there's a larger question.
What happens when producing the answer is no longer the hardest part of the assignment?
We have spent decades teaching students how to find answers. AI is getting remarkably good at producing them. The competitive advantage is shifting toward asking better questions, evaluating competing answers and deciding what to do when the answer isn't obvious.
That requires judgment.
And judgment is difficult to teach in a lecture.
Learning by Doing
The influencer marketing course described above was built at Queen Mary University of London using its Active Curriculum for Excellence, or ACE, framework. The framework emphasizes active participation, experiential learning, collaboration, self-paced work and practical application.
What interests me isn't the framework itself. It is what happened when AI was placed inside it.
Students didn't learn about AI by sitting through demonstrations. They had to use it.
They created scripts. They built synthetic media. They developed target audiences. They established performance measures. They tested ideas against actual marketing objectives.
Then they had to explain themselves.
That's where the learning got interesting.
One group, for example, developed a campaign for a beverage brand. They wanted to increase the brand's reach among adults aged 18 to 34. They established measurable goals around TikTok engagement, studied what their target customers valued and built an AI influencer around those findings.
Creating the video was only part of the assignment. The students had to decide whether their strategy made sense. They had to evaluate the cultural and ethical implications of what they created. They produced different scripts and compared them against their objectives.
Then their classmates got involved.
Students presented work at different stages and received structured feedback from their peers and instructor. They critiqued one another's campaigns using the same standards that would eventually be applied to their own work.
Think about what happened there.
AI made production easier. The course made thinking harder.
That's a combination worth paying attention to.
The Wrong Fight
I understand why educators worry about AI.
If I can type a question into a machine and receive a competent 1,500-word response thirty seconds later, asking a student to produce a 1,500-word response may no longer tell me what I think it tells me.
We can respond by trying to keep AI outside the classroom. There are circumstances where that makes sense. Students still need to learn how to think and write independently.
But prohibition alone strikes me as a losing strategy.
These tools aren't going away. Our graduates will enter organizations where AI is increasingly embedded in how work gets done. Telling students they can't use it throughout their education and then sending them into workplaces where they're expected to use it intelligently seems an odd form of preparation.
The harder work is redesigning learning so that AI can assist with production without outsourcing thinking.
That changes the instructor's job.
If AI can generate the first answer, perhaps the assignment should begin with that answer.
What's wrong with it?
What assumptions did it make?
What evidence would change your mind?
What ethical problem did the machine miss?
Would you stake your reputation on its recommendation?
Now we have something worth discussing.
Judgment Has Always Been the Scarce Resource
This matters well beyond higher education.
I've spent much of my career around senior executives and boards. The leaders I've admired most were rarely distinguished by having more information than everyone else in the room.
They knew what mattered.
They could sit with incomplete information, listen to competing arguments, recognize what they didn't know and eventually make a decision.
Sometimes they were wrong. Good judgment doesn't eliminate mistakes. It improves the quality of the thinking that precedes them.
AI gives leaders access to extraordinary amounts of information and analytical power. I suspect those capabilities will soon become commonplace.
That makes the human part more valuable, not less.
Can you distinguish confidence from competence?
Can you recognize bias, including your own?
Can you tell when an elegant analysis rests on a foolish assumption?
Can you ask the question nobody else thought to ask?
Can you make a decision when the data refuse to cooperate?
Those are leadership questions. They are also educational ones.
What Should We Be Teaching?
The experience at Queen Mary offers one useful answer.
Give students access to powerful tools. Then put them in situations where using those tools is only the beginning of the work.
Make them create something.
Make them defend it.
Expose their thinking to people who disagree with them.
Give them feedback and require them to improve the work.
Ask them to consider consequences that don't fit neatly into a spreadsheet.
That looks a lot more like the world they'll encounter after graduation.
It also changes the conversation about AI in education. The question becomes less about whether students used AI and more about what intellectual work remained theirs.
That's a much harder standard. I think it's also a better one.
The Part AI Can't Do for Us
I have no idea what the dominant AI tools will look like five years from now. I'm reasonably confident they'll make today's versions look primitive.
That doesn't worry me nearly as much as graduating people who have learned to use powerful technology without developing the judgment to question it.
The students in that marketing classroom didn't need another lecture about artificial intelligence. They needed an opportunity to use it, test it, argue about it and discover where its capabilities ended and their responsibility began. That is what the course gave them.
Business schools should pay attention.
So should CEOs.
AI will keep getting better at producing answers. Faster answers. More polished answers. Answers delivered with enough confidence to make bad ideas sound remarkably sensible.
Our job is to develop people capable of deciding which answers deserve to be believed.
Technology changes the mechanics of work.
Judgment determines the quality of the outcome.
Rob Andrews
Chairman & Chief Executive Officer
Celebrating 28 years of Executive Search, Executive Coaching & Culture Shaping Excellence
Direct: 713.489.9724/ Mobile: 713.301.6130
1800 St. James Place, Suite 510, Houston, TX, 77056
