After the spread of generative tools, the barrier to many tasks that once required specialized training has clearly dropped. Writing a piece of copy, building a prototype, processing a batch of data—all of these can now be done quickly with the help of tools. This convenience brings a side effect worth noting: users may simply accept the tool's output without asking further about the principles behind it.
What Does a Lower Barrier Bring
What tools lower is mainly the operational barrier, not the judgment barrier. Take data analysis as an example: tools can quickly produce charts and conclusions, but whether the data's definitions are consistent, whether the sample is representative, and under what conditions the conclusions hold—these questions still need to be answered by people.
Therefore, the divide brought by the spread of tools often lies not in who can use the tools, but in who can keep asking questions after the tool produces its output.
A Common Tendency: Skipping Understanding
When tools can directly provide usable results, the motivation to learn is easily weakened. Users gradually get used to "take the result first, think about the reason later," and may even treat the tool's output as the conclusion itself. In the short term, efficiency improves noticeably, but when encountering new problems that the tools cannot cover, those lacking foundational understanding often find it hard to keep moving forward.
This tendency has a greater impact in fields that require long-term accumulation. In work such as basic research, algorithm design, and system architecture, judgment comes from an understanding of the underlying logic, not from proficiency in operating tools.
What It Means for Companies
When introducing AI tools, companies usually focus on efficiency gains and easily overlook the corresponding changes in personnel capabilities. A more pragmatic approach is to make the tool's role clear: whether it handles repetitive steps or judgment-based steps. The former can be largely replaced, while the latter should retain a step for human confirmation.
Another approach is to retain the requirement for explanation. No matter who produces the result, they should be required to explain the basis and assumptions. This may seem to add workload, but in fact it effectively prevents the decline of judgment.
A Balanced Approach
- Treat tools as a starting point rather than an endpoint: after the output, still do a check and make a judgment.
- Retain foundational training: the more a role relies on tools, the more it needs to understand the basic principles behind them.
- Distinguish scenarios: for steps with low tolerance for error (public release, compliance, finance), insist on manual review.
- Encourage questions: treat "why is it like this" as part of the process, not an extra burden.
Tools will continue to get stronger—that much is certain. What truly requires long-term investment is whether the people using the tools can maintain independent judgment.