The fastest way to create resistance to artificial intelligence is to treat it as a technology rollout. Employees are not only judging the software. They are judging the leaders who selected it, the business case those leaders make, what they say about jobs, and what they avoid saying.
The AI conversation has become a leadership test
On September 1, 2026, HR Executive reported that employee discussion of AI had moved from largely positive to predominantly negative. The article was based on Glassdoor Economic Research published August 27 and analyzing U.S. employee reviews.
Glassdoor found that mentions of AI increased 240 percent from May 2025 to May 2026. In 2019, 81 percent of reviews mentioning AI were positive. In 2026, 43 percent were positive, 4 percent were mixed, and 53 percent were negative.
The signal is not that every employee opposes AI. They do not. The important signal is that employees increasingly connect workplace AI with leadership judgment, job security, burnout, training, surveillance, and whether the company still understands its own work.
The criticism was not limited to fear of replacement. Twenty percent of negative comments raised replacement concerns. Others objected to being forced to use AI, saw it distracting the company from its core business, questioned surveillance, or believed leaders expected unrealistic productivity gains.
Positive comments were more concentrated. Employees were more likely to support AI when they saw a commercial benefit or received useful tools, training, and support.
Glassdoor also found a strong relationship between AI sentiment and opinions of senior management. Only 23 percent of AI-critical reviewers approved of the CEO, compared with 86 percent of AI-positive reviewers. That is correlation, not proof that AI caused the difference. Employee reviews are also self-selected. But the size of the gap should get an executive team's attention.
Workers expect leaders to avoid the difficult conversation
A separate survey adds to the concern. LiveCareer surveyed 900 U.S. workers in July 2026 about trust during financial difficulty, layoffs, AI-driven change, reskilling, and employee well-being.
Seventy-one percent believed their employer would either downplay AI's effect on jobs or avoid discussing it until employees experienced the consequences. Seventy percent believed leadership would be unlikely to invest in retraining or reskilling people whose roles were affected. Eighty percent said they rarely or never trust leadership to put employee interests ahead of profit when the two conflict.
Those findings measure employee perceptions. They do not prove that every employer will act that way. The survey was also commissioned by a resume-services company. Still, perceptions affect whether employees listen, participate, share concerns, and believe the next announcement.
A leader cannot communicate only after every answer is final. By then, employees may have filled the silence with their own explanation.
Five leadership mistakes that turn AI into a trust problem
Announcing the tool without the business problem
Employees hear what the technology can do, but not which customer, quality, cost, speed, or workload problem the company is solving.
Avoiding the job-impact question
If leaders do not know whether roles will change, they should say that. People handle an incomplete answer better than a misleading one.
Requiring adoption without preparation
A license and login are not training. Employees need examples, decision boundaries, data rules, quality checks, and a path for reporting mistakes.
Giving managers a script instead of authority
Frontline managers receive the questions senior leadership avoids. They need information, authority, and communication skill.
Counting usage instead of business value
Logins and prompt counts are activity. They do not prove better quality, lower rework, faster cycles, better service, or controlled risk.
What leaders should do before the next AI announcement
Define the operating decision
State the business problem, the work that will change, the work that will not change, and the result the organization expects. Separate verified facts from estimates and open questions.
Clarify human accountability
Identify which decisions remain with people, who reviews AI output, what evidence is required, and who is responsible when the technology is wrong. AI can support work. It does not own the decision.
Prepare managers before employees ask
Managers should understand the purpose, limits, job impact, timeline, and support available. Give them a method for escalating questions they cannot answer. Do not reward confident guessing.
Build a listening loop
Ask employees where the tool improves work, where it creates rework, and where it creates risk. Then show what changed because of the feedback. Listening without visible follow-through will not repair trust.
Measure business results and leadership behavior
Track the operating outcomes the tool was supposed to improve. Also track whether leaders communicate clearly, listen, adjust, provide direction, and follow through. If the technology is improving while trust is collapsing, the rollout is not healthy.
Where CheckPoint 360° can help
CheckPoint 360° is a multi-rater leadership feedback and development tool. It compares a leader's self-evaluation with observed-behavior feedback from people who work with that leader, including supervisors, peers, and direct reports.
For AI-related change, the issue is not whether the tool can produce an AI-readiness score. It cannot. The value is identifying how the leader is experienced in behaviors that matter during change, including communicating effectively, listening to others, instilling trust, providing direction, adjusting to circumstances, and facilitating team success.
The report can help a leader identify a small number of development priorities, build an Individual Development Plan, and create accountability for follow-through. A later comparison survey may be used when appropriate to examine changes in observed behavior.
CheckPoint 360° should not be used as a hiring predictor, a stand-alone employment decision, or a guarantee that behavior or business results will improve.
Where Everything DiSC® on Catalyst™ can help
Everything DiSC® on Catalyst™ gives teams and leaders a practical language for behavioral preferences and tendencies. Selected applications allow participants to revisit their profiles, understand colleagues, and prepare for real interactions.
During an AI rollout, people respond differently to uncertainty, pace, conflict, detail, and change. DiSC can help leaders adapt how they communicate, recognize automatic reactions, and handle working-style differences without turning people into labels.
Everything DiSC supports development. It does not determine job fit, predict performance, or belong in pre-employment screening. A DiSC style is not an excuse for ineffective behavior.
Use the right tool for the right problem
| Business need | Useful contribution | Important limit |
|---|---|---|
| Leader behavior during change | CheckPoint 360° can show how the leader is experienced and focus a development plan. | It does not measure AI readiness or guarantee improvement. |
| Communication and working relationships | Everything DiSC® can create a shared language and help people adapt their approach. | It does not determine job fit or predict performance. |
| AI governance and operating controls | Policies, technical review, security, data controls, training, and human accountability are required. | A workforce assessment cannot replace technical or legal controls. |
The goal is not to assess people because the organization bought AI. The goal is to identify the leadership and communication problem, then use the right evidence and development method.
Measure whether the AI rollout is actually working
A responsible measurement plan begins before the rollout. Choose a small number of operating and people measures that the change can reasonably influence.
Operating measures may include cycle time, error rates, rework, customer complaints, service quality, output quality, or time spent on low-value tasks. People measures may include training completion, confidence using the tool, reported barriers, manager communication, unwanted-turnover signals, and documented follow-through on employee feedback.
Do not claim an assessment caused an operating result. Look for patterns, compare results with the baseline, account for other changes, and use human judgment.
The decision is not AI versus people
The stronger decision is whether leaders introduce AI with a clear business purpose, honest communication, defined human accountability, practical support, and a willingness to change course when the evidence says the rollout is not working.
Technology may change the work. Leadership determines whether employees understand the change, trust the process, and know what responsible performance now requires.
If an AI rollout is exposing gaps in trust, communication, or management consistency, define the development problem before choosing a tool. The Assessment Company® can help determine whether CheckPoint 360° or Everything DiSC® on Catalyst™ fits the need, and where another solution is required.