People say “don’t use AI” because they’re reacting to real risks that show up when AI is used without guardrails. The biggest concern is accuracy: AI can sound confident while giving incorrect details, mixing up sources, or inventing specifics. That’s frustrating in everyday tasks, but it becomes a serious issue in areas like health, legal, finance, customer support, and product claims where mistakes can harm someone or create liability.
Another reason is privacy and data security. When sensitive information (customer data, internal documents, pricing, passwords, medical details) is pasted into a tool, it may be stored, logged, or reviewed depending on settings and provider policies. Some teams also worry about intellectual property—AI outputs can blur ownership and may unintentionally mirror copyrighted material or proprietary text patterns.
There’s also a human side. Overreliance can weaken critical thinking, writing, and problem-solving skills, especially when AI becomes the default instead of a helper. And in the workplace, “don’t use AI” can reflect a boundary issue: the expectation to be always-on, producing more faster, with less recovery time. AI can create pressure to respond instantly and expand workload unless time and communication limits are set.
A practical middle path is to treat AI like a draft assistant, not a decision-maker. Keep private information out of chats, verify important facts with trusted sources, and define when AI is allowed (and when it isn’t). Setting time boundaries matters too—using AI to speed up tasks shouldn’t automatically increase the number of tasks that get assigned.
For a simple, time-based way to set limits—especially when saying “no” is hard—see the checklist here: https://estallos.com/guide-say-no-with-ai-checklist-time-boundaries/.
Decide which tasks you’ll use AI for, cap the time you’ll spend per request, and require human review for anything customer-facing or high-stakes. Communicate those rules early so speed gains don’t turn into unlimited availability.
Leave a comment