Most Workplaces Lack an AI Policy—Three in Ten Workers Use It Daily

Employee AI use is running ahead of workplace rules in the United States. SurveyMonkey’s Q3 results, published August 19, put daily use at 29%—21% multiple times a day and 8% once a day—while 55% of workers said their employer had no official AI-use policy.

The findings come from a July poll of 1,686 U.S. workers and students, as detailed in HR Dive’s independent coverage. The responses measure people’s use and awareness of employer rules rather than auditing company policies, but they expose a clear mismatch: routine use is widespread while formal guidance remains absent for a majority.

Daily use is different from occasional experimentation

Another 34% of workers use AI occasionally—a few times a week or less often—and 37% never use it at work. Those groups should not be collapsed into one adoption category: daily users have recurring workflows, occasional users make narrower choices about when to use the technology, and non-users may have little direct experience with it.

Daily users apply AI to research, brainstorming, messages, documents and data analysis. These activities carry materially different risks. Brainstorming with nonsensitive information creates a different exposure from entering client records into an external system, relying on generated analysis or sending unreviewed text to a customer.

Among AI users, 68% said the technology saved time, rising to 84% among daily users. Yet only 6% turn to AI first when encountering a task; 58% try to solve the problem before using it. The combination indicates substantial integration into work without uniform dependence.

The policy gap requires four distinct decisions

A policy that merely labels AI required, optional or prohibited leaves the most consequential questions unanswered. Beyond the 55% with no official policy, 34% described AI as optional at their workplace, 6% said it was required and 5% said it was completely prohibited.

The results support a policy-priority matrix based on task risk rather than a blanket judgment about the technology:

• Privacy and security: identify approved tools and specify which client, employee, financial or confidential information must not be entered. Privacy or security was the leading hesitation among both regular and occasional users.
• Accuracy and accountability: define which outputs require verification and who remains responsible for the finished work. Analysis, consequential decisions and external communications warrant tighter review than low-stakes ideation.
• Junior-worker supervision: state when a manager must review AI-assisted work and when an employee should complete the underlying task independently to develop judgment.
• Approved use cases: distinguish permitted research or drafting from restricted uses such as personnel decisions, legal conclusions, sensitive-data processing and unreviewed customer-facing material.

This ordering is an editorial synthesis of the uses and concerns in the results, not a ranking created by the researchers. It separates risks that require different controls: an employer can permit a low-stakes use without approving every tool, input or output.

Junior roles expose the limits of blanket rules

Workers remain divided over AI use in junior positions. In the Q3 breakdown, 42% supported use with clear guidelines and limitations, 14% supported use only with supervision and approval, 35% favored prohibition, and 9% supported unrestricted use.

The caution predates the latest release. SurveyMonkey’s preceding quarterly survey, published May 12, showed that 70% of workers preferred AI to be optional at an employer, while 57% thought companies should discourage its use in junior or entry-level positions.

The unresolved choice is whether restrictions should follow seniority or task risk. A seniority-based prohibition may preserve opportunities to learn foundational work, but it can also block supervised, low-risk uses. A task-based policy can allow drafting or research under review while reserving sensitive analysis, final judgments and external delivery for accountable human oversight.

Frequent users feel more secure than occasional users

Attitudes toward AI and job security change sharply with frequency of use. Among daily users, 27% felt more secure because of AI and 22% felt less secure. The balance reversed among weekly users, at 10% more secure against 25% less secure, and widened among people using AI once a month or less, at 3% against 24%.

The same divide appears in views of the broader labor market. Daily users were nearly split between optimism and pessimism, while weekly, occasional and non-users leaned negative. An employer policy written around enthusiastic daily users could therefore miss the uncertainty experienced by colleagues who use AI less often.

These figures do not establish that frequent use causes confidence. Workers with greater autonomy, stronger employer support or jobs that benefit readily from AI may be more likely to become daily users in the first place.

AI is also influencing career choices: 23% had changed or considered changing their target industry, 32% had changed or reconsidered which skills to develop, and 24% had changed or reconsidered the jobs or companies they were pursuing. Considering a change is not the same as completing one, but the figures connect workplace governance to longer-term questions about training, evaluation and career development.

The August findings establish a policy gap, not the effectiveness of any particular response. They do not reveal how many employers are drafting rules or whether restrictions improve accuracy, security or worker development. The next test is whether companies define approved tools, protected data, review duties and supervised uses while preserving the distinction between daily, occasional and non-users.

Also read:

• Why Breakfast is Must Include in Your Daily Routine
• Why AI Hasn’t Replaced Programmers — and Probably Won’t Replace Most Knowledge Workers
• How to Use ChatGPT Projects for a Startup Without Mixing Files, Instructions or Memory