Borrowed Competence: When AI Masks Your True Skill

Imagine this scenario: A marketing manager beamed as she presented the new campaign copy to her client. It was polished, engaging, and perfectly hit all the keywords. The client nodded, impressed, then asked, "Can you walk me through the strategic rationale for emphasizing 'agility' over 'speed' in this particular segment?" Sarah paused, her smile faltering. The AI had generated the copy; she had tweaked it for tone, but the deep strategic underpinning was something she hadn't fully grasped or vetted. The words were there, but the understanding wasn't. She had produced excellent work, but she couldn't defend it.

This scenario illustrates what I call Borrowed Competence: the ability to produce work that is objectively good, sometimes even excellent, thanks to AI, yet the underlying competence to create, judge, or defend that work independently isn't truly yours. It's a skill gap masked by technology, and it often only becomes apparent when the AI makes an error, or a nuanced challenge requires genuine human insight. This isn't an established scientific term, but rather a concept I've coined to describe a crucial challenge emerging with AI adoption in small and mid-sized businesses.

Why Borrowed Competence Differs from Plain Overconfidence

Borrowed Competence isn't just about being overly confident in your abilities; it's a specific kind of skill degradation and blind spot amplified by AI. Here's why it's different:

• Delayed Failure: Unlike traditional overconfidence, where mistakes might be immediate, Borrowed Competence often leads to delayed failures. Everything looks fine for weeks or even months. The AI produces quality output consistently, reinforcing a false sense of mastery, until a novel problem, an unexpected client question, or a subtle AI hallucination occurs.
• Exposure at Critical Moments: The true cost of Borrowed Competence hits precisely when the AI is wrong, or when a client asks a hard, probing follow-up that requires you to defend the nuanced decisions behind the work. You've shipped it, but you lack the foundational understanding to articulate why it's good, or how you'd adjust it under new constraints.
• Eroding Corrective Feedback: Getting truly good at any skill requires a continuous loop of creation, feedback, and correction. AI, by design, often smooths over the friction that produces correction. It fills in the gaps, polishes rough edges, and quickly presents a near-perfect solution. This removes the opportunities for you to struggle, make mistakes, and learn from those missteps—the very process that improves your error-spotting and critical judgment.
• Uneven Impact: AI helps most the people who already possess a strong understanding of their domain and can discern a truly good answer from a merely plausible one. They leverage AI to accelerate and enhance. It hurts most the people who lack this core judgment, as AI output appears authoritative, making it harder for them to identify subtle inaccuracies or suboptimal approaches.

The Pilot's Paradox: A Timeless Analogy for the AI Era

The airline industry faced a strikingly similar challenge in the 1980s with the advent of advanced autopilots. As automation became more sophisticated, pilots found routine flying incredibly easy. Autopilot made flights smooth, precise, and nearly flawless. However, an insidious side effect emerged: pilots, while highly trained, became less adept at the rare, critical moments when manual intervention was genuinely needed. They lost practice and muscle memory in handling complex emergencies because the automation was too good at preventing them.

There were documented incidents where highly skilled pilots struggled to regain control during unexpected failures, simply because they hadn't manually flown a plane under duress in years. The industry recognized this "pilot's paradox." Their solution? Mandatory manual flying hours in simulators and in the air, along with clear protocols for when the autopilot must be disengaged. They built in deliberate practice to maintain fundamental skills. For knowledge work empowered by AI, we currently have no equivalent built-in system of checks and balances.

What Businesses Can Learn: Practical Steps for Mitigating Borrowed Competence

Just as the airline industry adapted, SMEs embracing AI must proactively guard against Borrowed Competence. Here are actionable strategies you can implement this week:

1. Only Delegate Work You Could Judge if You Had To: If you can't assess the quality, accuracy, or strategic fit of an AI's output yourself, you haven't delegated; you've gambled. Ensure that any task you outsource to AI is within your realm of understanding for evaluation. This aligns with What Top Companies Do Differently With AI – they maintain oversight.
2. Guess First, Then Prompt: Before asking AI for an answer or a draft, take a moment to formulate your own expected answer or outline. Write it down. Then, generate with AI. The gap between your preliminary guess and the AI's output is a real-time measurement of your current skill level and where AI is truly augmenting you. This isn't just about getting an answer; it's about continuously learning. It's a form of Just-in-Time Learning: Skill Development in the AI Era applied to your daily workflow.
3. Decide How Hard You'll Check BEFORE You Generate: It's human nature: once you see a polished, coherent output from AI, you're less likely to scrutinize it critically. Before you even type your prompt, decide on the stakes and how rigorously you'll verify the information. For low-stakes content, a quick glance might suffice. For critical decisions or client-facing deliverables, plan for a thorough fact-check and strategic review before you see the AI's answer.
4. Keep Doing Some Work Manually (On Purpose) in Your Core Skill: Don't let AI completely automate away your fundamental expertise. Dedicate specific time each week or month to perform core tasks in your domain manually. This isn't nostalgia; it's how you maintain the ability to spot a bad answer, understand nuances the AI might miss, and ensure your own judgment remains sharp. This kind of deliberate practice is essential for 3 Skills to Future-Proof Your Career in the AI Era.
5. Know the Difference Between "I Can Produce This" and "I Can Defend This": Be explicit with clients, your team, and even yourself about the nature of the work. If you used AI, acknowledge that the production was assisted. More importantly, be honest about whether you can fully defend the strategic depth or factual accuracy without AI's continued help. Transparency builds trust and manages expectations.
6. For Teams: AI Moves the Bottleneck from Producing to Reviewing: If your team leverages AI for content creation, code generation, or data analysis, understand that the throughput of production will increase dramatically. However, the bottleneck often shifts to review and validation. If you don't adequately staff and train for this critical review process, you've simply built a faster pipeline for mistakes, not for quality output. This highlights the importance of Process Over Tools: Be Dogmatic About Your Approach.

Ultimately, addressing Borrowed Competence isn't about shunning AI; it's about integrating it wisely. It's a discipline choice, not a tool choice. Your commitment to maintaining human judgment and critical thinking alongside powerful AI tools will define your business's true intelligence and resilience.

Frequently Asked Questions

What is Borrowed Competence?

Borrowed Competence describes the situation where an individual can produce high-quality work using AI tools, but lacks the underlying independent skill or knowledge to fully create, judge, or defend that work without AI assistance. This often leads to delayed failures when true expertise is required.

How does AI contribute to Borrowed Competence?

AI contributes by making it easier to generate polished output quickly, effectively masking skill gaps and reducing opportunities for self-correction and genuine learning. It can prevent users from engaging in the struggle and feedback loops necessary for deep skill development, making them reliant on the AI's output without necessarily understanding its foundations.

How can businesses prevent Borrowed Competence among employees?

Businesses can prevent Borrowed Competence by encouraging critical review of AI output, mandating some manual work to maintain core skills, implementing strategies like "guess first, then prompt," and training employees to understand the limitations and potential biases of AI tools. Investing in human oversight and skill development alongside AI adoption is key.

Is Borrowed Competence the same as overconfidence?

No, Borrowed Competence is distinct from simple overconfidence. Overconfidence implies an inflated belief in one's existing skills. Borrowed Competence, however, describes a situation where the ability to produce (thanks to AI) outstrips the ability to judge or defend that output, leading to delayed discovery of skill gaps rather than immediate errors from overestimation. The failure is often only apparent when the AI falters or deeper human insight is required.