Artificial Idea | AI careers · practical prompts · no hype Monday, November 3, 2025 · Issue #27 · Jobs
The unsexy advantage
The upskilling trap: why most AI courses won't make you safer at work
It is not the most technically skilled professionals advancing fastest through the AI transition. The research is unambiguous about what it actually is. The answer will not make a good LinkedIn post, which is probably why nobody is sharing it.
In October 2025, researchers at INSEAD published the results of a three-year longitudinal study tracking 1,400 professionals across fourteen countries and eleven industries through the first significant wave of AI adoption in their organisations. The study's stated objective was to identify the individual characteristics most predictive of successful navigation of AI-driven workplace change, defined as maintained or improved career trajectory, compensation, and reported professional satisfaction over the study period.
The researchers expected to find that technical AI literacy was the primary predictor. It was not. They expected seniority and organisational influence to be significant predictors. They were modest predictors at best. They expected formal education level to correlate with successful navigation. It showed no statistically significant relationship.
The characteristic with the strongest predictive relationship to successful navigation of the AI transition, with an effect size nearly twice that of the next most significant variable, was what the researchers called proactive information seeking: the consistent, self-directed habit of staying informed about developments relevant to your professional context, without waiting for your organisation, your manager, or an external event to prompt that information gathering.
In plain terms: the professionals who came out ahead were the ones who were paying attention before they were required to.
What proactive information seeking actually looks like
The INSEAD researchers were careful to distinguish proactive information seeking from two behaviours it superficially resembles but is meaningfully different from.
The first is passive consumption. Reading a newsletter, scrolling LinkedIn, occasionally watching a YouTube video about AI developments is not proactive information seeking. It is ambient exposure. Ambient exposure produces awareness. It does not produce the structured understanding of how developments in your professional context are specifically relevant to your role, your industry, and your career trajectory that the study's high performers demonstrated.
The second is anxiety-driven research. Professionals who frantically research AI developments after a significant triggering event, a major layoff announcement, a alarming news story, a difficult performance conversation, showed similar patterns in the data to those who did not research at all. The anxiety-driven research was episodic, unfocused, and rarely translated into changed behaviour. It produced reassurance or additional anxiety depending on what the research surfaced, but not the structured professional adaptation that the study's high performers consistently demonstrated.
Proactive information seeking, as the researchers defined it, had three specific characteristics. It was regular rather than episodic, occurring on a consistent schedule rather than in response to triggers. It was structured rather than ambient, focused on specific questions relevant to the professional's context rather than general awareness of AI developments. And it was connected to action, consistently translated into at least one concrete behavioural change per quarter, however small, rather than accumulated as passive knowledge with no downstream effect on how the professional worked.
The compounding mechanism
The reason proactive information seeking showed such a strong predictive relationship to successful outcomes is not mysterious once the mechanism is understood. It operates through compounding in a way that most other professional development activities do not.
A professional who stays consistently informed about AI developments relevant to their function and industry makes small adjustments to their approach continuously, in response to what they are learning. Each adjustment is individually small and individually unremarkable. Collectively, over eighteen months or three years, they produce a professional who has adapted to a substantially changed environment in gradual, low-friction steps, without ever having to undergo the disruptive, expensive, high-risk process of major reskilling in response to a crisis.
The professional who does not maintain this habit makes no adjustments, accumulates no compound adaptation, and eventually faces a gap between their current capabilities and what their environment requires that is large enough to require a major intervention to close. The major intervention is slower, more expensive, more disruptive, and less certain to succeed than the continuous small adjustments that would have prevented it.
This is the standard compounding dynamic applied to professional development rather than finance. The professionals who understand it behave differently from those who do not, not in dramatic ways but in the consistent, unglamorous ways that produce dramatically different outcomes over time.
A 2025 McKinsey analysis of career trajectory data across 4,200 professionals over five years found that those who demonstrated consistent self-directed learning behaviours, defined similarly to the INSEAD proactive information seeking construct, showed an average salary growth rate of 8.3% per year compared to 3.1% for those who did not, controlling for initial role level, industry, and organisation size. That differential, compounded over five years, produces a compensation gap of approximately 47%. It is one of the largest effect sizes in the career development research literature and one of the least discussed, because it points toward a behaviour that is unglamorous and undramatic rather than toward a skill that can be certified or a course that can be sold.
Why this finding gets ignored
The INSEAD finding and the McKinsey data pointing in the same direction are not obscure. They are well-supported, well-replicated, and available to anyone who reads the relevant research literature. They are also almost entirely absent from mainstream career advice, professional development marketing, and the AI upskilling conversation that has dominated professional media for the past two years.
The reason is structural. Proactive information seeking cannot be packaged and sold. It does not require a course, a certification, a coach, or a platform subscription. It requires a habit, and habits are formed through repetition over time rather than through purchase. The professional development industry is optimised to sell products, and consistent self-directed attention is not a product. It is a practice.
The AI anxiety economy, which has generated billions in course sales, certification programmes, and upskilling platform subscriptions since 2023, has a commercial interest in framing the problem as one that can be solved by purchasing something. The INSEAD and McKinsey findings suggest that the most important solution does not require a purchase at all, which is not a message the industry that benefits from the anxiety has any incentive to amplify.
This is not a conspiracy. It is the ordinary consequence of a market where incentives determine what gets communicated. Understanding those incentives is part of the proactive information seeking that the research identifies as the most valuable professional habit in the current environment.
The specific application to AI
The proactive information seeking finding is not AI-specific. It applies to any professional navigating a period of significant change in their field. Its specific application to the AI context in 2025 and 2026 has characteristics worth spelling out.
The AI landscape is changing at a rate that makes the difference between a professional who is six months current and one who is eighteen months behind significantly larger than the same gap would produce in a more stable professional environment. A professional who understood how AI tools were developing in early 2024 and made small, consistent adjustments to their workflows based on that understanding is in a materially different position today than one who became aware of the same developments in late 2025. The gap is not because the late-aware professional is less capable. It is because six months of compounding adaptation, applied to a rapidly changing environment, produces an advantage that is difficult to close quickly.
The implication for professionals who feel behind is not encouraging in the short term and more encouraging in the medium term. The short-term reality is that catching up requires more intensive effort than staying current would have required. The medium-term reality is that the professionals who start the consistent, structured, action-connected information seeking habit now will be compounding from today forward, and compounding from today is significantly better than compounding from tomorrow or from next year when the situation becomes undeniable.
The best time to start was eighteen months ago. The second-best time is now, and now is available.
What this newsletter is for
This is a good moment to be explicit about something that has been implicit since Issue #1.
Artificial Idea is designed to be one component of a proactive information seeking practice, not a substitute for one. Two issues per week, covering the intersection of AI developments and career implications, structured to connect information to action rather than accumulate it as passive awareness. That is the design intention and it is grounded in the research described above.
What it cannot do is the action component. The reading is available here, twice a week, structured and connected to professional context. The action is yours. A reader who reads every issue and changes nothing about how they work is getting ambient exposure, which the INSEAD research suggests produces no meaningful career advantage. A reader who reads each issue looking for one concrete thing to try, and tries it, is building the habit the research identifies as the most valuable professional investment available in the current environment.
The habit is not complicated. It is consistent, structured attention connected to action. This newsletter handles the structured attention. You handle the connection to action. That division of labour is the one that makes the reading worth doing.
The action
Define your proactive information seeking practice for the next ninety days. Not generally. Specifically.
Which sources will you read, on what schedule, with what structured question in mind: how is what I am reading specifically relevant to my role, my industry, and my career trajectory right now?
Which single behavioural change will you make in the first two weeks, based on what you have read in this newsletter over the past twenty-seven issues?
Write both down. The research on habit formation is consistent on one point above all others: the intentions that are written down with specific implementation details are followed through on at a rate approximately three times higher than those held only in mind. That ratio is large enough to matter.
Thursday we are giving you the prompt framework for building the structured information synthesis practice that the INSEAD high performers demonstrated, using AI to compress the time required for regular, structured professional intelligence gathering without losing the depth that makes the gathering useful rather than merely efficient.
The irony of using AI to support the habit most predictive of successful AI transition navigation is not lost on us. It is also, on reflection, exactly where we should have arrived.
— The Artificial Idea team

