AI can produce a convincing answer in seconds. But when everyone has access to the same tools, knowing what to question, what to trust and when to change your mind becomes more important. Brendan and Ross catch up with Dr Jason Fox at IT Nation 2026 to talk about AI, discernment, uncertainty and staying relevant when nobody quite knows what comes next.
Watch the Conversation
Watch the conversation on YouTube
The Short Version
As AI makes sophisticated answers easier to access, human judgement becomes more important, not less. Dr Jason Fox argues that businesses need people who can question AI outputs, seek opposing views, notice unusual signals and stay open to changing their minds as new information appears.
What’s Covered
Six Things Worth Taking Away
AI can sound convincing even when it is wrong. Fluency and expertise are not the same thing.
As access to AI grows, people still need the judgement to decide what is relevant, credible and worth acting on.
Doubt can improve thinking. Asking “Where am I wrong?” helps expose assumptions that might otherwise go unchallenged.
Leaders need to give people enough certainty to work while leaving room to question assumptions as conditions change.
Not every useful signal appears neatly in the data. The odd comment or recurring customer complaint may be worth investigating.
Best practice has its place, but following the same playbook as everyone else can make it harder to spot what should come next.
AI is making access to information and apparent expertise much easier.
Jason’s concern is that an answer can sound convincing without necessarily being correct. That becomes especially difficult when the person reading the response does not already know enough about the subject to recognise the gaps.
As he puts it:
“Large language models are very impressive, but particularly if you don’t have subject matter expertise.”
That makes discernment increasingly important.
It is the ability to look at an answer and ask whether it actually makes sense in this situation. What assumptions sit behind it? What might be missing? Does it fit what we know about the business? What would someone with a different view say?
For businesses adopting AI, access to the tool is only part of the equation. Someone still has to judge the result.
Jason uses the term “midwit trap” to describe what he calls the performance of sophistication.
An answer becomes longer. There are more complicated words. There are frameworks and explanations. The sheer amount of information makes it feel impressive.
AI is particularly good at producing this kind of material.
The risk is that we walk away feeling as though we understand something more deeply than we actually do.
Jason argues there is often more honesty in being able to say:
“I don’t really know. But here’s the questions we’re working on.”
That matters in business because polished output can create a false sense of certainty.
A lengthy AI-generated strategy is not automatically better than a short document that clearly explains what is known, what is uncertain and what still needs to be tested.
Probably not.
The conversation touches on a familiar problem with generative AI. The direction you give it can shape the answer you receive.
Ask why your idea is good and you can get an excellent explanation of why it is good. Keep doing that and it becomes easy to build an increasingly convincing argument around your original assumption.
Ross describes using different AI models to argue opposing sides of an issue rather than asking one system to keep developing his existing position.
Jason sees value in that kind of resistance.
“Doubt is evidence of thinking.”
Instead of only asking AI to develop an idea further, businesses can use it to challenge their reasoning.
Try questions such as:
The point is not to create endless debate. It is to give an idea enough resistance to see whether it still holds up.
Jason describes the current period as a time “betwixt worlds”.
Businesses have spent years improving systems, practices and operating models built for a relatively stable environment. AI and other emerging technologies are now changing some of the assumptions underneath them.
The next stable model is not yet clear.
That creates an uncomfortable position for leaders, particularly those who built their careers by being the person with the answer.
Jason suggests leaders should instead create what he calls “islands of coherence”.
Give people enough stability and clarity to do their work, but do not make the organisation so certain of its existing model that it stops paying attention.
That might mean holding a working hypothesis rather than pretending to have a permanent answer.
It might mean saying:
“We think this is the right direction based on what we know today. Here’s what we’re watching, and here’s what would make us reconsider.”
Jason uses his self-described “wizard” persona to make a serious point.
Powerful technology should be used, not ignored. But it also requires perspective.
Businesses do not need to respond to AI by avoiding it altogether. At the same time, being swept up in every new capability can make it harder to judge whether the technology is actually helping.
A useful question is not simply:
“What can AI do?”
It is:
“Does using AI here make this work better?”
That means keeping enough human understanding in the process to recognise when the technology is helping, when it is not, and when the way the business works may need to change.
Good data matters. But relying only on established data can also reinforce what the business already believes.
Jason talks about developing an “acuity for anomaly”.
The people speaking to customers or doing the work every day are often the first to notice something unusual.
A customer keeps raising an odd complaint.
The same unexpected question appears in several conversations.
Someone in operations notices a workaround becoming common.
One of those signals may look insignificant. Several appearing over time may suggest something worth investigating.
The difficulty is that businesses do not always have a formal place to capture these faint signals.
Jason’s suggestion is surprisingly simple: give people opportunities to compare notes.
That could be a regular team conversation, an informal weekly catch-up or a longer lunch each month where people can ask:
“What are you noticing?”
Eventually, something that first looked like an outlier may turn into a pattern.
Best practice is useful when you need a proven way to do something.
But there is a trade-off.
If everyone follows the same best practice, everyone gradually starts doing similar things in similar ways.
The conversation contrasts “best practice” with “next practice”.
That does not mean throwing out processes that work. It means continuing to ask whether the established approach still makes sense as the context changes.
AI makes that question particularly relevant.
A process designed around what people and technology could do several years ago may not be the process you would design today.
Instead of only asking:
“How does everyone else do this?”
Businesses can also ask:
“Knowing what is possible now, how would we design this if we started again?”
One of the more unusual parts of the conversation begins with poor Wi-Fi.
Jason describes spending time writing in a cabin with very limited internet access and noticing a different quality of attention.
Without an immediate answer a click away, he had to stay with the uncertainty of a problem for longer.
He connects this to poet John Keats’ idea of “negative capability”: the ability to remain with complexity, ambiguity and paradox without immediately reaching for a neat solution.
That is difficult when an answer is always available.
The spare moment that once involved thinking can now become a search, a notification or a question to an AI assistant.
There is obvious value in having answers close at hand. But instant access to an answer and thinking deeply about a problem are not necessarily the same thing.
Sometimes the business may benefit from sitting with the question a little longer.
Jason closes the conversation by returning to why many people entered the technology industry in the first place.
They saw ways technology could make people’s lives better.
That leads to a distinction between profit and value.
Financial return matters, but Jason argues that not every form of value is easily captured in a number.
Value can also appear in the relationships a business builds, the experience customers have and whether technology genuinely improves the work people do.
That is a useful test for AI as well.
The question should not stop at:
“Can we use AI here?”
It should continue to:
“What becomes better if we do?”
Before Accepting the AI Answer
AI can help you develop an idea quickly. Before acting on it, spend some time trying to disprove it. That can expose assumptions, missing information and alternative explanations that the first answer did not include.
Ask:
A Practical Model for AI-Assisted Decisions
About the Guest
Philosopher, complexity practitioner and self-described wizard
Jason’s work draws on philosophy, living systems science and complexity. In this conversation, he explores what wisdom and discernment look like when AI can give almost anyone access to sophisticated answers, and why staying relevant requires the willingness to question what worked in the past.
Common Questions
Full Episode Transcript
Want to move from experimenting with AI to putting it to work across your business?
IT Strategy
IT Consulting
Data Services
IT Infrastructure Projects
IT Procurement
Business Continuity Planning
Virtual CIO
Cloud Migration Services
Managed IT Services
IT Support Services
Managed Cloud Services
Managed SharePoint Services
Unified Communications
Connectivity
Offshore Staff Outsourcing (BPO)
Managed Security Services
ACSC Essential 8
Cybersecurity Risk Assessment
Cloud Backup and Disaster Recovery
Cyber Security Connect and Protect Program
Automation
Power BI
Careers
Awards and Certifications
Management Team
Case Studies
Sell Your IT Business
Our Commitment to the Environment
CSAT Charity Donations
Media Kit
Sydney
