# Exclusive: Workers wary of AI performance reviews-66% trust tools less, 85% want humans making final hiring decisions

> AI speeds reviews and training, but workers still trust managers more and want humans making the final call. Use AI to help, be upfront, and keep judgment human.

- Type: AI news
- Page: https://completeaitraining.com/news/exclusive-workers-wary-of-ai-performance-reviews-66-trust/
- Published: 2025-10-21
- Updated: 2025-10-21
- Author: Jeroen Erne
- Jobs: Management
- Reading time: 3 min

### How employees view AI management tools - and what managers should do next

AI can speed up reviews, training, and onboarding. But employees are still cautious - especially about fairness and final decisions.

In a new Software Finder survey of 1,006 employees, 21% say an AI tool has evaluated their performance. Yet 23% felt misjudged by an AI system, and 66% trust AI tools less than human-led performance management. Only 12% trust them more; 22% see them about the same.

There's optimism, though: 59% feel positive about AI's growing role at work. Still, only one-in-three believe AI improves their ability to succeed; 24% think it threatens it, and 18% think it does both. Eighty-five percent want a human manager or HR to make the final call in hiring.

#### Generational signal, not a green light

Gen Z is the most open to AI judging performance (14% trust it more than managers). But they're also the least comfortable being hired or fired by AI (9%). Millennials are the most comfortable with AI in those decisions - and it's still just 11%.

#### What employees think is most accurate

Seven-in-10 say human feedback best reflects their value. About one-in-four prefer a mix of AI assessment plus human input. Millennials are the most supportive of that hybrid approach (about 28%).

#### Learning and onboarding: use AI, but don't make it cold

Twenty-six percent say AI makes learning at work easier; 42% say it makes learning feel impersonal; 32% say both. Gen X is most likely to say AI helps (28%) and least likely to say it makes learning feel robotic (38%).

- 46% have used an AI training or learning tool.
- 55% prefer hybrid learning - AI plus human support.
- AI-enabled onboarding can backfire: 46% felt less connected to the company; 43% felt no difference.

### Manager playbook: deploy AI without losing trust

#### 1) Keep AI in an assistant role for performance

- Use AI to summarize goals, trends, and examples - not to assign final ratings.
- Require human calibration and a documented rationale for every rating and pay decision.
- Offer an appeal process when AI-informed data influences outcomes.
- Ban AI-only "productivity surveillance" scores from compensation and promotion decisions.

#### 2) Hiring and promotions: human makes the final call

- Make it policy: 85% of employees want a human or HR to decide. Do not delegate final decisions to AI.
- Audit tools for job relevance and adverse impact. Re-test after model or job changes.
- Tell candidates and employees when AI is used, how it's used, and what data it uses.
- Provide a non-AI alternative or accommodations where needed.

Helpful references: [EEOC guidance on AI and employment](https://www.eeoc.gov/laws/guidance/eeoc-issues-new-technical-assistance-use-software-algorithms-and-artificial?ref=completeaitraining.com), [NIST AI Risk Management Framework](https://www.nist.gov/itl/ai-risk-management-framework?ref=completeaitraining.com).

#### 3) Learning and onboarding: go hybrid by default

- Pair AI modules with live coaching, manager check-ins, and peer cohorts.
- Use AI for personalization and spaced practice; use humans for context, judgment, and culture.
- In onboarding, add mentors, scheduled 1:1s, and "day 7/30" pulse checks to offset the 46% connection drop.
- Track completion, application on the job, and manager-observed behavior change - not just quiz scores.

#### 4) Transparency isn't optional

Trust erodes when people feel kept in the dark: 23% say their company isn't transparent about AI at work, and 34% aren't sure.

- Publish an AI-at-work notice: tools in use, purposes, data sources, safeguards, and human oversight.
- Provide a feedback and appeal channel for AI-assisted decisions.
- Maintain an AI tool register with owners, risks, mitigations, and review dates.

#### 5) Measure what matters

- Performance: fairness gaps by team/role, calibration variance, appeal rates and outcomes.
- Hiring: selection rate differences, time-to-fill, quality-of-hire, candidate satisfaction.
- Learning: adoption, skill gains, manager ratings of on-the-job use, retention during and after onboarding.
- Trust: quarterly pulse on perceived fairness, clarity, and confidence in AI-assisted processes.

#### 30-60-90 day rollout (practical)

- Days 1-30: Inventory AI in use; pause high-risk use cases (firing, discipline, promotions). Draft an AI policy with human-in-the-loop requirements.
- Days 31-60: Pilot hybrid training and AI-assisted reviews with two teams. Run bias and impact tests. Set up an appeal path and publish your AI notice.
- Days 61-90: Expand based on results. Add dashboards for fairness and trust metrics. Schedule quarterly audits.

### Bottom line for managers

Employees are open to AI - as long as humans stay accountable, decisions are explainable, and the process feels fair. Use AI to reduce grunt work and improve consistency. Keep judgment, context, and final decisions with managers.

#### Skill up your team

If you're setting standards or training managers on AI-assisted performance, hiring, or L&D, explore curated options here: [AI courses by job role](https://completeaitraining.com/courses-by-job/).

---

This is the Markdown form of https://completeaitraining.com/news/exclusive-workers-wary-of-ai-performance-reviews-66-trust/. The whole site is mapped at https://completeaitraining.com/llms.txt.
