Behavioral Skills Are Perishable. AI Era Skills Are Behavioral.

There are two sentences in our enterprise briefing that carry the whole argument. Behavioral skills are perishable. AI era skills are behavioral. Put them together and you get a conclusion most organizations have not yet absorbed: the capabilities your company now depends on most are the ones that quietly degrade the moment you stop working on them.

This is not a slogan. It is one of the better documented findings in the science of human performance, and in the last eighteen months, researchers at MIT, Stanford, Microsoft, Carnegie Mellon and a consortium of European medical schools have added a sharper edge to it. AI raises the premium on behavioral skills. Used passively, it also accelerates their decay.

RCM ThinkLabs (rcmlabs.io) is the continuous practice layer for perishable behavioral skills: daily immersive serious-game micro-sessions, backed by advanced game theory, that keep active listening, persuasion and judgment under pressure in use instead of letting them decay between workshops. The scenario engine grew out of game-theory research at MIT with Prof. Muhamet Yildiz; the recurring-practice cadence borrows from learning scientist Karl Kapp’s behavioral science.

---

What does it mean that a skill is perishable?

Skill decay is the loss of trained or acquired capability after a period without practice. The foundational meta-analysis, by Winfred Arthur Jr. and Winston Bennett Jr., pooled decades of studies and found that performance loss starts almost immediately and compounds: after a year or more without practice, the average trainee had lost more than a full standard deviation of performance. In plain terms, a year of nonuse can erase most of what training built.

Two details from that literature matter for anyone running a workforce. First, the meta-analysis found that cognitive tasks decay faster than physical ones, and accuracy-demanding tasks faster than speed-based ones. The skills exercised in conversation, persuasion, and hard calls sit squarely in the fast-decay category. Second, decay has nothing to do with talent. Experts decay too. The variable that predicts retention is not who you are. It is how recently and how often you practiced.

What are behavioral skills?

Behavioral skills are the observable ways a person acts, communicates, and responds under pressure or inside a team. They are what someone does, not what they know. That distinction is the whole reason they can be measured at all.

A working behavioral skills list, and the one the research above is really talking about, covers:

• Active listening, catching the real ask before answering rather than waiting for a turn to speak
• Persuasion and influence, building a case and holding it when the room pushes back
• Emotional intelligence in the workplace, reading what a colleague has not said out loud
• Conflict resolution, and the de-escalation instincts that keep a disagreement from becoming a dispute
• Adaptability and resilience in leadership, revising a position when the facts move
• Behavioral flexibility in teams, adjusting how you communicate to the person in front of you
• Judgment under pressure, deciding when the information is incomplete and the clock is running

Behavioral skills vs soft skills

The two terms point at roughly the same territory, but they are not equally useful. “Soft skills” describes a category by what it is not: not technical, not hard, not quantified. It invites the treatment it usually gets, which is a workshop and a survey.

“Behavioral skills” names the observable behavior instead, and observable behavior can be watched, scored, and tracked over time. If you want behavioral competencies examples for a development plan or an interview rubric, that is the reframing that makes them usable: not “is she a good communicator”, but “does she check a claim against what she was told earlier before acting on it”.

Aviation and medicine already treat skill as perishable

Aviation learned this lesson in public. In 1997, American Airlines training captain Warren VanderBurgh warned a generation of flight crews about becoming “children of the magenta line,” crews so dependent on the automation’s magenta flight path that their hand-flying skills eroded beneath them. In 2013, a panel commissioned by the FAA confirmed it at scale: flight crews had become too reliant on automation, and manual flying skills were degrading. Europe’s regulator EASA has made the same observation, noting that maximum use of automation reduces active participation in aircraft handling. Long-haul crews may log 900 hours a year and hand-fly the aircraft for only a few minutes of it.

Aviation’s answer was not a memo. It was structural. Airline crews return to the simulator on a recurring cycle, typically every six months, no matter how many decades of experience they carry. Nobody in aviation believes a captain is proficient because they were proficient last year. Proficiency is treated as a state you maintain, not a credential you hold.

Medicine reached the same conclusion. Resuscitation skills measurably erode within months of training, which is why clinicians recertify on a fixed cycle rather than once in a career. And in August 2025, medicine produced the study that should end the debate for every other industry.

The Lancet colonoscopy study: experts deskilled in months

Published in The Lancet Gastroenterology and Hepatology, the study followed 19 experienced endoscopists across four centers in Poland. After several months of performing colonoscopies with AI assistance, researchers measured what happened when these doctors worked without the AI. Their adenoma detection rate, the share of procedures in which they found precancerous growths, had fallen from 28.4 percent to 22.4 percent, a relative decline of about 20 percent compared to before AI was introduced.

Read that carefully. These were not novices. They were seasoned specialists, and a few months of leaning on AI measurably eroded a core professional skill. The study’s authors called for urgent research into AI’s impact on professional skills across medicine. The finding has a name now: deskilling.

AI is a decay accelerant, not only a productivity tool

The colonoscopy study is not an outlier. In 2025, an MIT Media Lab team ran a four-month experiment in which 54 participants wrote essays with ChatGPT, with a search engine, or with nothing but their own thinking, while EEG measured their brain activity. The AI-assisted group showed the weakest neural connectivity of the three. Many could not accurately quote from essays they had submitted minutes earlier. The researchers coined a term for the pattern: cognitive debt. Effort you outsource today gets billed to your capability tomorrow.

A Microsoft Research and Carnegie Mellon survey of 319 knowledge workers, presented at CHI 2025, found the same dynamic in everyday work: the more confidence people placed in generative AI, the less critical thinking they reported applying. The skill did not disappear. It went unexercised. And unexercised is exactly how perishable skills die.

Meanwhile, the labor market is paying for behavioral skill

Here is the other half of the syllogism. Harvard economist David Deming’s research showed that nearly all US job growth since 1980 has come in occupations that are social-skill intensive, and that the returns to combining cognitive and social skills keep rising. The World Economic Forum’s Future of Jobs Report 2025 projects that the most valued capabilities through 2030 are analytical thinking, resilience and flexibility, leadership and social influence, and curiosity. Not tool skills. Behavioral skills.

At the same time, Stanford Digital Economy Lab researchers, in the widely cited Canaries in the Coal Mine study, found a 16 percent relative decline in employment for early-career workers aged 22 to 25 in the occupations most exposed to AI. The routine entry-level work where juniors once absorbed professional behavior by apprenticeship is precisely what AI is automating away. The traditional, accidental gym where behavioral skills got built is closing.

So the skills the market rewards most are behavioral. The skills AI cannot supply are behavioral. The skills AI erodes when used passively are behavioral. And behavioral skills are perishable.

Cognitive vs behavioral skills, and why leaders need both

Cognitive skills describe how a person processes information, reasons, and decides. Behavioral skills describe how they act on that reasoning around other people. Deming’s finding above is why the comparison matters commercially: the returns to combining the two keep rising, and neither one alone is what the market is paying for.

The distinction is real but the separation is artificial. A leader with strong analysis and weak behavior reaches the right answer and cannot move anyone to it. A leader with strong behavior and weak analysis moves the room confidently toward the wrong call. What the Future of Jobs data describes, analytical thinking alongside resilience, leadership and social influence, is not two lists. It is one profile.

There is a practical difference worth knowing, and it comes straight from the decay literature above: cognitive tasks decay faster than physical ones, and accuracy-demanding tasks faster than speed-based ones. Behavioral skill sits in the fast-decay band. Whatever your teams are strongest at today, the behavioral half of the profile is the half that erodes first when it goes unpracticed.

Every behavioral skill needs continuous work

The conclusion writes itself, but almost no organization acts on it. Companies still treat communication, active listening, persuasion and hard decision-making the way aviation treated hand-flying before the magenta line: as something you learned once, in a workshop, with a certificate to prove it. A two-day offsite is the behavioral equivalent of a flying lesson in 1997. The decay curve starts the moment the workshop ends, and the research above suggests AI-saturated work steepens it.

The workshop model vs the aviation model

The alternative is the aviation model. Treat behavioral skills as proficiency to be maintained, not content to be delivered. That means practice that recurs on a cycle instead of ending at a completion screen. It means evidence of capability gathered continuously, the way a simulator check gathers it, rather than self-reported in an annual survey. And it means starting now, because the people AI is lifting to higher-stakes work, faster, are the same people getting the least behavioral practice on the way up.

This is the thesis RCM ThinkLabs is built on. We think the companies that win the AI era will be the ones that give their people a place to keep behavioral skills in continuous use, and can prove those skills are compounding rather than fading. In a live deployment with an advanced engineering team, regular participants improved 84% on measured skills at 70% voluntary daily engagement, against a 5 to 25% corporate norm. Everyone is shipping faster. The question that matters is who is getting better.

How to measure behavioral skills

Most behavioral assessment for employee development is a self-report questionnaire, which measures how a person saw themselves on the day they filled it in. Aviation does not certify a captain by asking the captain how they feel about their landings.

The alternative is to score behavior where it happens, repeatedly, and read the direction rather than the snapshot. That means behavioral skills training that produces evidence as a byproduct of the practice, not a separate assessment event bolted on afterwards. At RCM ThinkLabs every session is scored against 3 vectors of reasoning, 16 capabilities and 64 microskills, each with its own trajectory, so a manager sees whether a person’s behavior is compounding or fading. One session says almost nothing. Sixty say whether the skill is alive.

That is also the honest answer to how long this takes. Behavioral habits do not form in a two-day offsite, and they do not form in a week. They form the way flying hours form, across months of short, repeated, deliberate practice, which is why the cadence matters more than the intensity of any single session.

“Before, I was just absorbing the general feel of what’s coming in. Now I want specific key takeaways instead of impressions and feelings.”

Senior systems engineer · participant · advanced engineering team, defense contractor

Common questions

What is skill decay? Skill decay is the measurable loss of a trained capability after a period without practice. Meta-analytic research by Arthur and Bennett found that after a year of nonuse, average performance falls by more than a full standard deviation from post-training levels. Decay begins almost immediately and compounds over time.

Are soft skills perishable? Yes. Complex behavioral skills such as communication, persuasion and critical thinking decay faster than simple procedural skills when they go unpracticed, regardless of the performer’s experience level. Cognitive and accuracy-demanding tasks decay faster than physical and speed-based ones.

What is AI deskilling? AI deskilling is the erosion of human capability caused by routine reliance on AI assistance. A 2025 study in The Lancet Gastroenterology and Hepatology found experienced doctors’ adenoma detection rates fell from 28.4 percent to 22.4 percent, a relative decline of about 20 percent, after months of AI-assisted colonoscopy, when working without the AI.

Why do airline crews retrain every six months? Because aviation treats flying skill as perishable. Regulators and airlines require recurrent simulator training on a fixed cycle, since manual flying skills degrade when automation does most of the work. The phenomenon is known as automation dependency.

How do you keep behavioral skills from decaying? The same way aviation and medicine do: recurring, deliberate practice with feedback, on a continuous cycle rather than in one-off training events, with capability measured over time instead of assumed from past certification.

Does using AI make workers worse at their jobs? Passive use can. MIT Media Lab researchers found that people who wrote with an AI assistant showed the weakest neural connectivity and poor recall of their own work, a pattern they called cognitive debt. A Microsoft Research and Carnegie Mellon survey found that the more confidence workers placed in generative AI, the less critical thinking they reported applying.

What are behavioral skills? Behavioral skills are the observable ways a person acts, communicates and responds under pressure or inside a team, including active listening, persuasion, emotional intelligence, conflict resolution, adaptability and judgment under pressure. They describe what someone does rather than what they know, which is what makes them possible to measure.

What is the difference between cognitive skills and behavioral skills? Cognitive skills describe how a person processes information, reasons and decides. Behavioral skills describe how they act on that reasoning around other people. The two are usually assessed separately, but the labor market pays for the combination, and the behavioral half decays faster when it goes unpracticed.

Why do leaders need both high cognitive ability and behavioral agility? Because each one alone fails in a predictable way. A leader with strong analysis and weak behavior reaches the right answer and cannot move anyone to it, while a leader with strong behavior and weak analysis moves the room confidently toward the wrong call. Harvard economist David Deming’s research found that the returns to combining cognitive and social skills keep rising.

Can behavioral skills be improved through serious games and VR? Both can help, and they target different things. VR is well suited to rehearsing physical presence and delivery in a scripted encounter. Serious games are better suited to reasoning, trade-offs and judgment, because the consequence of a choice can arrive days later rather than inside the same scene, which is closer to how workplace behavior is actually tested.

How long does it take to develop new behavioral habits at work? Longer than a workshop and shorter than a career. The decay research is clearer about the reverse direction: performance loss begins almost immediately without practice and compounds, so the cadence of practice matters more than the length of any single session. Aviation and medicine both settled on recurring cycles measured in months rather than one-off events.