AQai analysed 19,077 first AQ assessments from 2020 to August 2026. Mental Flexibility has held steady, but Unlearning, the ability to deliberately let go of what you know, has been falling. Here is what the data shows, what it does not, and the one question to ask your organisation.
Across 19,077 first AQ assessments taken between 2020 and 1 August 2026, one ability has been slipping while almost everything around it has held steady: Unlearning. Unlearning is the ability to deliberately let go of what you already know and reassess it. Its close partner, Mental Flexibility, has not moved. That leaves organisations with a practical question: what are you still doing because it worked?
Both abilities sit within AQ-Ability, part of Adaptability Intelligence (AQ). Unlike personality-style preferences, they can be developed, and they can also fade when they are not used. This article sets out what our data shows, what it does not show, and where to start.
Mental Flexibility is the ability to accept, appreciate and embrace competing demands or problems. Unlearning is the ability to intentionally let go of previous knowledge, and to reassess based on new and old data. The simplest way to hold them apart: Mental Flexibility is how many things you can hold at once. Unlearning is what you are willing to put down.
They are two halves of the same system, the intake and the release. They are also easy to mislabel. Unlearning is not forgetting. Forgetting is passive. Unlearning is a choice to re-examine something you believe, so that it no longer governs the decision. It is also not being persuaded. Somebody else can give you a good argument, but only you can reassess your own position.
This is why "I am open to new ideas" tells you little. It usually describes Mental Flexibility. A better test of Unlearning is the last thing you changed your mind about, not the last new idea you heard.
Unlearning scores have fallen consistently since 2022, while Mental Flexibility has stayed flat. We looked at first assessments only, so we are describing a population and not the effect of our own development work. We trend from 2022 onward, and 2026 is a part year, running to 1 August.
Average Unlearning scores by year of first assessment were 68.5 in 2022, 66.3 in 2023, 65.1 in 2024, 63.4 in 2025 and 64.3 in 2026 so far. Average Mental Flexibility scores were 60.4, 62.6, 62.0, 61.8 and 61.3 over the same years. One note on reading these: the two abilities are scored against different bands, so it is the movement over time that is comparable, not the two score levels against each other.
Another way to see it is through the bands we use to describe a result as low, medium or high. These bands are fixed and have not changed, which is what makes drift visible. In 2022 about 3 in 10 people landed in the low Unlearning band. Today it is about 4 in 10.
No. The decline holds when we compare like with like. Keeping job level, age group and gender constant across 16,289 first assessments from 2022 onward, Unlearning still falls by about 1.2 points a year. Mental Flexibility does not change in any meaningful way.
The stronger evidence is what did not move. Grit, mindset, resilience, hope, work stress and overall AQ were all steady across the same period. If our sample had drifted, or working people were simply having a harder time, several of these would have moved together. Only Unlearning did.
We should be straightforward about size: the effect is small by conventional standards. It matters because it is isolated, consistent year on year, and it compounds. The middle of the population has moved furthest, with a typical score falling from 70 to 63, while the very top has barely changed. That is why the typical person today would have ranked in the bottom 40% in 2022. It is also happening everywhere, with every job level from entry-level to executive showing a decline.
The data does not support that. Work stress averages 50.7 for people with low Unlearning and 47.0 for people with high Unlearning, and the relationship between the two is close to zero. Uncertainty about change tells a different story: 38.5 for the low group against 23.6 for the high group.
People higher in Unlearning also report more support from their company and their team. Environment appears to be doing something here, just not through workload. It points toward a different question: is it safe to be wrong in front of colleagues?
The two abilities travel together, so most people sit in a similar place on both. The more interesting people are the ones who do not. The group with high Mental Flexibility and low Unlearning has grown from about 1 person in 40 in 2022 to about 1 in 17 now. Its mirror image, low flexibility with high Unlearning, has thinned from about 1 in 16 to about 1 in 38. The group high on both has shrunk from 23.3% to 17.4%.
Put simply, we used to see more people who let go too easily and struggled to hold complexity. We now see more people who handle complexity well and will not let go of anything. They generate alternatives, then keep the original anyway.
When we compare this group with people who are high on both, grit is almost identical (84.7 against 86.0). Nobody here is lazy. What differs is their relationship with change. Uncertainty about change is higher (33.9 against 23.3), and resilience is lower (70.5 against 80.7). So the question worth asking is not "what else could you try?" but "is the next change feeling riskier than the last one?"
A caution applies. We measured these people once, at one moment. We can see what travels together, and we cannot say what causes what. These are positions people occupy today, not types of people.
Neither Mental Flexibility nor Unlearning is a virtue on its own. Both are contextual. A low position serves a determined mission where deviation carries real risk: emergency services, the military, aviation and regulated compliance work. Turning off contrarian thought is the job, and doing it well protects people. A high position serves dynamic conditions, ambiguous problems and work that rewards experiment, especially where the proven process is quietly going out of date.
So the question is not which is better. It is whether the position fits the context, and for how long that context holds. Bloodletting was standard medical practice for centuries. Somebody had to be mentally flexible about it, and somebody had to unlearn it. A stance that served an industry for a generation may now serve it for a year.
One further trap: somebody may be fully open to an alternative and still unable to act on it, because of hierarchy, resourcing or deadline pressure. A score cannot tell willing-and-unable apart from unwilling. The only way to find out is to ask.
We do not know, and we would rather say so than guess. Four explanations are worth holding open, and our data cannot rule any of them in or out:
What we can say is what it is not. It is not workload, and it is not a change in who we assessed. We tested both.
Not through effort and persistence. Unlearning has far more to do with resilience, hope and how somebody handles not knowing than it does with grit or how busy they are. The right door is curiosity and a willingness to take risk, building the capacity to bounce back from experiments and failures, rather than the capacity to push through.
A practical starting point is to ask better questions, in one-to-ones, team meetings and development conversations:
Because the shift is in the broad middle and not at the edges, treat this as work for the whole population, not a remedial problem for a few strugglers. Scoped that way, you will miss almost all of it.
We measured people once, at a single moment, so we can see what travels together but not what causes what. These are not measures of intelligence, capability or performance. And we are not suggesting that any group, including senior leaders, is the problem: the people with the most accumulated success simply have the most to let go of.
No. Forgetting is passive and involuntary, while Unlearning is a deliberate choice to re-examine something you believe. In AQ terms, it is not losing the knowledge. It is deciding that the knowledge no longer governs the decision.
We do not know. Across 19,077 first AQ assessments from 2020 to August 2026, the decline holds when job level, age group and gender are controlled, and it is not explained by workload. Possible explanations include a post-pandemic return to normal, change fatigue, too much to unlearn at once, and environments that make it unsafe to be wrong, but our data cannot confirm any of them.
No. A low reading describes where someone sits today, not the type of person they are, and Unlearning is an ability that can be developed. Low can also be the right place to be in work where deviation carries real risk, such as aviation or emergency services. A score also cannot tell somebody who is unwilling from somebody who is willing but unable.
Focus on resilience, hope and comfort with uncertainty rather than effort. In AQai's data, Unlearning travels most closely with resilience, and has much less to do with grit or work stress. Useful questions include "What would have to be true for it to be safe to stop doing this?" and "What is our organisation still doing because it worked?"
Mental Flexibility is the ability to accept, appreciate and embrace competing demands or problems: how many things you can hold at once. Unlearning is the ability to intentionally let go of previous knowledge and reassess based on new and old data: what you are willing to put down. Both are part of AQ-Ability, they correlate, and they develop through different routes.
Start with one question in your next team conversation: what is our organisation still doing because it worked? If you want to see where you or your people sit on Mental Flexibility, Unlearning and the other drivers of adaptability, explore the AQme assessment or read more about the AQ model.
Copyright adaptai Ltd. All rights reserved. AQ® and AQai® are registered trademarks of adaptai Ltd.
.png){kind=small}
.png)
