This article was originally published on my blog and translated into English. For the latest version and future updates, please visit the original post: https://jaketao.com/language/en/ai-exposure-gap-between-people
In 2026, generative AI (GenAI) gradually evolved and took shape as so-called "agents," entering everyone's lives at an astonishing pace (regardless of the specific concept, this article collectively refers to them as "AI" for the sake of simplicity). From content creation and information retrieval to daily office work, an increasing number of professionals are turning to AI to enhance efficiency and support decision-making processes. The integration of AI has enabled the rapid transformation of abstract concepts into tangible realities. Software development is a relevant example here: in the past, many people had a wealth of creative ideas and product concepts, but were unable to bring them to life due to a lack of programming skills. However, the advent of AI-powered coding has effectively lowered the barriers to entry across the entire process, from requirements specification to code generation and deployment. In today's digital landscape, developers have the ability to rapidly create product prototypes and entire applications based solely on conceptual ideas.
Consequently, disparities between individuals have also widened. For some, AI serves as an accelerator of efficiency; for others, it's more like a shortcut that bypasses the process of learning and accumulation. AI has the potential to empower individuals to achieve tasks that were previously challenging, but it also brings to light issues that were once difficult to identify.
Ownership: The strong tend to become more productive, while the weak become less active.
Laziness is an inherent aspect of human nature, which is why ownership is of particular importance. In the AI era, what is truly scarce is not the act of getting things done itself, but rather judgment, critical thinking, and a sense of responsibility for the results. Those who lack ownership often merely "check off tasks."
To illustrate, consider the following example. Your superior requests that you conduct industry research or compile a set of materials. Some individuals may opt to delegate this task to artificial intelligence, refining it slightly before generating a comprehensive report with a well-designed structure, sophisticated formatting, and professional language. The document appears to be of a high caliber, potentially surpassing the quality of many individuals' own work. This can lead to a tendency to skim through the material and submit it without thorough review.
However, it is important to assess whether such an outcome is truly valuable. While the content generated by AI is often accurate, it frequently lacks the critical element of thought. AI is not equipped to determine the rationale behind a particular need for information, identify the challenges currently being faced by the team, or ascertain the true company priorities. While the answers it provides may be comprehensive, objective, and logically sound, they could also be a mere pile of "correct nonsense." This is due to the fact that AI is particularly adept at answering questions but lacks the capacity to effectively solve problems. This observation aligns with the concept of "talkers," as discussed in the article "Stay Away from Those Who Just Talk." While many people are impressed by the results generated by AI, which are often considered to be flawless or to exceed expectations, the true value of these results is yet to be determined.
This impression arises because the content remains at the informational level without addressing the actual problem. A person who truly takes ownership of a task does not first think about "how to complete it" when it is assigned, but rather "why it needs to be done." To determine the appropriate angle for conducting research and identifying areas requiring the most attention, it is essential to thoughtfully consider these questions. They will break the problem down into multiple parts, develop their own analytical framework, and then use AI to verify, supplement, and challenge their own judgments. This process is significantly more time-consuming than simply querying AI, and the value of the final output is entirely different.
- What is the rationale behind the boss's need for this information?
- What are the team's current challenges?
- What decisions will this information ultimately influence?
- Which details are essential and which are actually irrelevant?
This dynamic has led to a noteworthy development in the AI era: the prosperous are becoming more productive, while the less fortunate are becoming more leisurely. The company has adopted a strong use of AI to transform its work methods, breaking down, optimizing, and enhancing tasks that were previously impossible to complete due to time and cost constraints. The advent of AI has not diminished their intellectual capacity; on the contrary, it has enabled them to delve more profoundly into complex issues. Conversely, underperformers often approach AI with a superficial mindset, perceiving it merely as a solution provider. While it may seem that they are completing more tasks, they are actually just reducing their own thought process. For these professionals, the time saved by AI merely allows them to complete work faster, shifting the responsibility of thinking to others.
AI has not reduced the workload; it has merely shifted the focus from the execution level to the thinking level. This will also become the standard for screening and weeding out talent in the next era.
While AI-generated work is indeed identifiable at first glance, this is not the primary concern.
Many people are perplexed by this phenomenon. I would like to inquire about the reason for my boss's displeasure upon reviewing the polished report that was generated by artificial intelligence. Could his concerns be related to my use of AI?
In most cases, the boss is not concerned with the authorship of the AI-generated content. He is upset because you did not dedicate yourself fully to the project.
To you, this might seem like a high-quality piece of work --- visually appealing, well-structured, logically sound, and even accompanied by charts and data analysis. However, even after devoting ten or twenty minutes to meticulously reviewing numerous pages, it can be challenging to identify truly valuable information, which is often obscured by a substantial volume of professional text. While AI has proven to be highly effective in generating content, it lacks the capacity to discern what is truly important or to determine which approaches are practical and actionable. Consequently, individuals with a deficiency in critical thinking often find themselves susceptible to pitfalls, such as equating ostentatious content with value and misapplying the concept of comprehensive expression to imply profound comprehension. The end result is a report that "looks professional" but is full of "nonsense."
- What is the conclusion?
- What are the reasons for this occurrence?
- Which risks pose the greatest threat?
- What are the next steps?
The "One Pager," a term coined by Amazon, is a common requirement in business projects and reports, emphasizing the need to convey information succinctly on a single page. If the core points cannot be distilled onto a single page, including additional content will not improve the document. This is analogous to the one-page limit for resumes.Additionally, when highly skilled individuals utilize AI, their deliverables often become more concise. They utilize artificial intelligence to gather information, validate hypotheses, and expand their thinking. However, when they present their findings to decision-makers, they often limit the information to just the most important conclusion.
As barriers are eliminated, disparities begin to manifest.
AI is rapidly transforming the landscape of knowledge and skill sets, but this evolution does not guarantee equitable outcomes. Instead, it is likely to exacerbate existing disparities in wealth and opportunity. Historically, the development of software necessitated the expertise of programmers, financial analysis required a background in finance, and the design of marketing strategies demanded marketing experience. In today's business world, artificial intelligence (AI) has become a valuable asset, allowing finance professionals to develop apps, product managers to write code, and engineers to quickly learn about investing, law, or even marketing. The cost of acquiring cross-disciplinary skills is decreasing. Could this imply that industry experience and professional expertise are no longer significant factors?
On the contrary, the evidence suggests that the opposite is true.
Historically, the disparities among individuals primarily stemmed from limitations in accessing information, knowledge, and skills. With the significant advancements in AI technology, these barriers have been lowered, allowing other factors to come to the fore. These include judgment, depth of understanding, and the quality of thinking. In summary, AI has reduced the competency gap but increased the cognitive gap. While AI can assist in coding, it lacks the capacity to evaluate the viability of a product for development. Similarly, while AI can aid in generating a business plan, it cannot ascertain market demand. Additionally, while AI can support in data analysis, it cannot determine the significance of data or the reliability of conclusions.
Take AI Coding, for example. In today's business world, professionals in fields such as product management, operations, and marketing can utilize tools like Claude Code and Codex to expedite the development of websites, systems, and the deployment of products to production environments. This is an exciting development. The true challenge, however, lies not in the construction of the product itself, but rather in the creation of a product that truly meets the needs of its intended users. While AI can assist in implementing features, it does not proactively question requirements. It aids in problem-solving, but does not indicate whether a problem is worth solving. While it provides answers, it rarely alerts users to potential flaws in the problem itself. Its effectiveness lies in its ability to provide answers to inquiries rather than defining the terms itself.
While AI coding can facilitate development, it does not inherently reduce the complexity of the system itself. The operational readiness of a demo is not necessarily indicative of its ability to support real users, and the long-term stability of a feature is not guaranteed by its initial implementation. From access control and data security to performance optimization, monitoring and alerts, disaster recovery, cost control, and compliance requirements --- not to mention the various edge cases and anomalies that may arise as the user base grows --- these issues won't automatically disappear just because AI is involved. Often, the most challenging aspect of software development is not the writing of code, but the anticipation of problems, the design of solutions, and the making of trade-offs under various constraints.
Despite the integration of artificial intelligence (AI), certain individuals will likely demonstrate heightened proficiency, while others may face challenges in identifying their needs and selecting the appropriate keywords to effectively convert into prompts for AI-driven processing.
AI will facilitate the initiation of many individuals, yet advanced expertise and comprehension will persist in their scarcity. As the number of professionals in a given field increases, the focus shifts from the question of whether one can do something, to whether one has the necessary understanding, and from whether one can build something, to whether one can build it correctly.
Soul
I have selected "soul" as my concluding point because, in the current era of artificial intelligence, its significance has increased to an unprecedented degree. Today's AI systems are designed to learn from and synthesize the knowledge and experience accumulated by humans over time. These systems can imitate, integrate, and generate, but they lack the capacity to possess their own values, beliefs, or aspirations. The true essence of a work is not found in the tool itself, but rather in the individual who wields it.
AI has the potential to make knowledge readily accessible and execution more efficient than ever before. However, it cannot replace the experience and insights accumulated by a professional over many years. It is essential to emphasize that a sense of responsibility, judgment, values, and accountability for outcomes is not only indispensable but will be significantly amplified.
Top comments (0)