Quick Article Summary

• Texas employers can use artificial intelligence in the workplace, but AI tools must be managed carefully because existing employment laws still apply to hiring, promotion, discipline, termination, disability accommodations, background checks, and employee monitoring.
• Employers should never allow AI to make final employment decisions without human review, because the business remains responsible for discriminatory, inaccurate, or unfair outcomes even when a third-party vendor provides the tool.
• Responsible workplace AI requires written policies, manager training, bias review, data protection practices, employee transparency, and periodic AI audits to make sure the technology supports good HR decisions instead of creating legal risk.

---

Why AI in the Workplace Is No Longer a Future Issue

Artificial intelligence is no longer something only large technology companies use. Small and mid-sized businesses are already using AI in everyday workplace decisions, sometimes without realizing it. An employer might use ChatGPT to help write a job description, use recruiting software to rank applicants, use payroll software that flags attendance patterns, use a scheduling tool that predicts staffing needs, or rely on productivity software that scores employee activity. These tools may seem administrative, but once they affect hiring, pay, discipline, promotion, termination, or access to employment opportunities, they become HR compliance issues.

For Texas employers, the central question is not whether AI can be useful. It can be. The real question is whether the employer can use AI without creating discrimination risk, privacy problems, wage and hour issues, employee relations problems, or defensibility issues in front of the EEOC, Department of Labor, Texas Workforce Commission, or a court. Business owners should treat AI like any other powerful workplace tool: useful when controlled, dangerous when left unchecked.

The most important principle is simple: AI may assist workplace decisions, but it should not replace human responsibility.

What AI Actually Means for Small and Mid-Sized Employers

When business owners hear “AI,” they often think of advanced robotics or complicated technology they will never use. In reality, workplace AI is much broader and more common. In an HR setting, AI may include software that drafts documents, screens resumes, evaluates candidate responses, predicts employee turnover, monitors productivity, schedules shifts, summarizes complaints, creates training materials, or identifies performance patterns.

For example, an employer using AI to draft a job posting is using AI in a relatively low-risk way, as long as the final posting is reviewed by a human for accuracy, legality, and professionalism. However, an employer using AI to automatically reject job applicants based on resume keywords is using AI in a much higher-risk way, because the tool may unintentionally screen out protected groups or qualified applicants whose resumes do not match the algorithm’s preferred pattern.

The level of risk depends on how the tool is used. AI used for brainstorming internal ideas is very different from AI used to decide who gets interviewed, who gets promoted, who gets written up, or who gets terminated. The closer AI gets to an employment decision, the more oversight the employer needs.

The IBM Principle: AI Should Augment Human Intelligence, Not Replace Human Responsibility

“A computer can never be held accountable, therefore a computer must never make a management decision.”

– IBM Training Manual, 1979

One of the strongest ways to think about workplace AI comes from IBM’s long-standing responsible AI position. IBM’s Principles for Trust and Transparency state that the purpose of AI is to augment human intelligence. That principle matters deeply in HR because employment decisions affect people’s income, reputation, career opportunities, and legal rights.

In plain language, a machine should not become the responsible party for a human employment decision. The employer remains responsible. The manager remains responsible. HR remains responsible. If an AI tool recommends rejecting a candidate, terminating an employee, denying a promotion, or flagging someone as a low performer, the business cannot simply say, “The system made the decision.”

That is not a defense. It is an admission that the company allowed a tool to make a decision without adequate oversight.

For HR purposes, the better rule is this: AI can assist, organize, summarize, and identify patterns, but a trained human must review the facts, apply policy, consider context, and make the final decision.

Federal Employment Laws Already Apply to AI

A common misconception is that workplace AI exists in a legal gray area because there is no single federal “AI employment law” covering all uses. That is the wrong way to look at it. Employers should assume that existing employment laws already apply when AI is used in employment decisions.

The EEOC has made clear that employers can violate federal discrimination laws when AI or algorithmic tools are used in employment decisions. The agency’s resource explaining the EEOC’s role in AI states that AI and other technologies may violate laws prohibiting discrimination when used in employment decisions involving protected characteristics such as race, color, religion, sex, national origin, disability, age, or genetic information.

This means Texas employers should evaluate AI tools under the same legal framework they use for traditional employment practices. A resume screening tool, assessment test, video interview analyzer, or employee scoring system may still be subject to federal anti-discrimination laws if it affects employment opportunities.

AI and Title VII: Disparate Impact Risks in Hiring and Promotion

Title VII prohibits employment discrimination based on race, color, religion, sex, and national origin. AI can create Title VII risk when a tool appears neutral but disproportionately screens out applicants or employees from a protected group.

The EEOC has specifically addressed this concern in its technical assistance on assessing adverse impact in software, algorithms, and artificial intelligence used in employment selection procedures, which explains that employers may be responsible when algorithmic decision-making tools create a selection rate that disadvantages individuals based on protected characteristics.

For business owners, this matters because a vendor’s promise that a tool is “objective” does not eliminate risk. A tool can be mathematically consistent and still legally problematic. For example, if an AI screening tool ranks candidates based on factors that correlate with age, race, national origin, or sex, the employer may face discrimination exposure even if no one intended to discriminate.

A Texas employer using AI in hiring should ask whether the tool has been tested for adverse impact, whether the employer understands the factors being measured, and whether the tool is actually job-related. If the vendor cannot explain how the tool works or how it has been validated, the employer should be cautious.

AI and Title VII: Disparate Impact Risks in Hiring and Promotion

Title VII prohibits employment discrimination based on race, color, religion, sex, and national origin. AI can create Title VII risk when a tool appears neutral but disproportionately screens out applicants or employees from a protected group.

The EEOC has specifically addressed this concern in its technical assistance on assessing adverse impact in software, algorithms, and artificial intelligence used in employment selection procedures, which explains that employers may be responsible when algorithmic decision-making tools create a selection rate that disadvantages individuals based on protected characteristics.

For business owners, this matters because a vendor’s promise that a tool is “objective” does not eliminate risk. A tool can be mathematically consistent and still legally problematic. For example, if an AI screening tool ranks candidates based on factors that correlate with age, race, national origin, or sex, the employer may face discrimination exposure even if no one intended to discriminate.

A Texas employer using AI in hiring should ask whether the tool has been tested for adverse impact, whether the employer understands the factors being measured, and whether the tool is actually job-related. If the vendor cannot explain how the tool works or how it has been validated, the employer should be cautious.

AI and the ADA: Disability Discrimination Is a Major Risk

One of the clearest areas of federal concern is disability discrimination. The EEOC’s guidance on artificial intelligence and the ADA warns that software, algorithms, and AI tools used to assess applicants or employees may disadvantage people with disabilities if the tools measure traits or abilities affected by a disability.

The Department of Justice has also warned that AI tools can cause disability discrimination in hiring. Its guidance on algorithms, artificial intelligence, and disability discrimination in hiring explains that tools used to screen, rank, or assess applicants may unlawfully screen out people with disabilities if they do not fairly measure the person’s ability to perform the job.

This can happen in practical ways. A video interview tool might score eye contact, facial expression, speech patterns, or response speed. That may disadvantage individuals with visual disabilities, speech impairments, neurological conditions, anxiety disorders, or other disabilities. A timed online assessment might screen out a qualified applicant who needed a reasonable accommodation. A personality test might penalize traits associated with a disability without actually measuring whether the person can perform the job.

The key HR takeaway is that employers must provide reasonable accommodations in the hiring process. If an applicant says they cannot complete an AI-based assessment because of a disability, the employer should not ignore the request or rely blindly on the tool. The employer should evaluate whether an alternative assessment is needed.

AI and Age Discrimination

AI can also create age discrimination risks. The Age Discrimination in Employment Act protects individuals who are 40 or older from employment discrimination. While AI may not directly ask for age, it may rely on factors that indirectly favor younger applicants or disadvantage older workers.

For example, a tool might prioritize recent graduation dates, certain technology keywords, social media activity, or short career histories. If those factors are not truly necessary for the job, they could create age-related risk. Employers should be especially careful with AI tools that rank candidates based on “culture fit,” “energy,” “digital native” traits, or other vague concepts that may hide age bias.

The safer approach is to focus AI-assisted screening on clearly job-related skills, certifications, experience, availability, and essential qualifications.

AI in Hiring: Where Employers Get Into Trouble

Hiring is one of the highest-risk areas for workplace AI because hiring decisions are heavily regulated and often involve large numbers of applicants. AI may be used to draft job ads, screen resumes, rank candidates, schedule interviews, analyze interview responses, conduct assessments, or recommend finalists.

The risk increases when AI is used to reduce a large applicant pool without meaningful human review. If the tool rejects qualified candidates based on flawed assumptions, biased training data, inaccessible testing methods, or irrelevant criteria, the employer may never know it lost strong candidates or created legal exposure.

Employers should avoid using AI as an automatic rejection machine. A better practice is to use AI as an administrative support tool while preserving human review for all meaningful employment decisions. For example, AI can help organize resumes by minimum qualifications, but a trained human should review borderline candidates, accommodation requests, and any rejection criteria that may disproportionately affect protected groups.

AI-Generated Job Descriptions and Job Postings

Using AI to draft job descriptions can be helpful, but it still requires HR review. AI may create job descriptions that sound polished but include legally risky language, unrealistic requirements, unnecessary physical demands, or vague duties that do not match the actual job.

A job description should accurately reflect the essential functions of the role. This matters for hiring, ADA accommodations, performance management, and termination decisions. If an AI tool exaggerates requirements or adds unnecessary qualifications, the employer may unintentionally screen out qualified candidates or weaken its position later in an accommodation dispute.

Employers should use AI-generated job descriptions as drafts only. HR or management should review the final version to confirm that the duties are accurate, the qualifications are necessary, and the language does not discourage protected groups from applying.

AI Resume Screening: A Practical Employer Standard

If an employer uses AI to screen resumes, it should create a written standard before the tool is used. That standard should identify the required qualifications, preferred qualifications, disqualifying factors, and human review process.

For example, a defensible resume screening process might state that the AI tool may identify applicants who meet minimum criteria such as required licenses, years of relevant experience, or specific technical skills. However, the final decision about who receives an interview should be reviewed by a human decision-maker who can evaluate context.

The employer should also avoid screening criteria that are not truly job-related. If a degree is not actually required to perform the job, the AI tool should not automatically reject applicants without one. If a certification is preferred but not required, the tool should not treat it as mandatory. If employment gaps are not relevant, the tool should not automatically penalize them.

AI Interview Tools and Video Analysis

AI interview tools create special risk because they may analyze speech, facial expressions, tone, word choice, response timing, or other factors that may not reliably predict job performance. These tools may also raise ADA concerns if they disadvantage applicants with disabilities.

If an employer uses AI-assisted interviewing, it should be transparent with candidates, provide accommodation options, and avoid relying solely on automated scores. The employer should also ask the vendor what factors the tool evaluates, whether the tool has been tested for bias, and whether the tool is validated for the specific job.

A practical standard is this: If the employer cannot explain what the AI interview tool is measuring, it should not rely on the tool to make hiring decisions.

Background Checks, AI Scores, and the FCRA

Some AI tools are not just hiring tools. They may also function as consumer reports or generate employment-related scores based on third-party data. When employers use consumer reports for hiring, promotion, retention, or reassignment, they must comply with the Fair Credit Reporting Act.

The Federal Trade Commission explains that employers using consumer reports for employment decisions must provide proper notice, obtain written authorization, and follow adverse action procedures when employment decisions are based on those reports.

The Consumer Financial Protection Bureau has also addressed modern screening tools in its circular on background dossiers and algorithmic scores for hiring, promotion, and other employment decisions. This is important because some employers may not realize that algorithmic reports, worker scores, or third-party employment data products may trigger FCRA obligations.

The practical point is simple: if a third-party tool provides a report, score, recommendation, or profile used to make an employment decision, the employer should evaluate whether FCRA rules apply before using it.

Can Employers Use AI for Discipline or Terminations?

Employers can use AI to identify patterns that may be relevant to discipline or termination, but they should not use AI as the final decision-maker. This is especially important when discipline could later be challenged through the Texas Workforce Commission, the EEOC, or litigation.

AI might flag repeated tardiness, productivity changes, unusual transaction patterns, customer complaints, or failure to complete required tasks. Those flags may be useful, but they are not proof by themselves. A manager or HR professional must verify the underlying facts.

For example, if an AI attendance tool flags an employee as unreliable, the employer should still review time records, approved leave, medical issues, schedule changes, system errors, and whether the policy was applied consistently. If an AI productivity tool identifies an employee as underperforming, the employer should verify whether the tool measured meaningful work or simply counted activity that may not reflect actual performance.

A termination based only on an unexplained AI score is weak. A termination based on verified facts, policy violations, documented warnings, and human review is much stronger.

AI and TWC Unemployment Claims

From a Texas Workforce Commission perspective, AI-generated information may help support a claim response, but it should not replace human testimony or documentation. If an employer terminates an employee for misconduct connected with the work, the employer must be prepared to explain what happened, what policy was violated, how the employee knew the rule, and why the conduct justified termination.

If the employer says, “The software determined the employee was a poor performer,” that is unlikely to be persuasive. If the employer says, “The software identified missed tasks, but management verified the records, reviewed the employee’s written expectations, issued warnings, and confirmed repeated failure to follow instructions,” the case becomes much stronger.

AI can help locate patterns. It cannot replace the employer’s burden to prove the facts.

AI Employee Monitoring and Privacy

Employee monitoring is another high-risk area. Many employers use software that tracks keystrokes, websites, GPS location, application activity, idle time, call metrics, or productivity scores. Some tools now use AI to interpret behavior and assign risk scores.

Employers may have legitimate reasons to monitor work activity, especially when employees work remotely, handle sensitive data, drive company vehicles, or work in regulated environments. However, monitoring should be transparent, limited to business purposes, and tied to a written policy.

Employers should avoid secretive or overly invasive monitoring. The more personal the data, the higher the risk. Tracking work activity during work hours is one thing. Collecting personal communications, private device data, off-duty activity, or sensitive biometric information is another.

The Department of Labor’s materials on artificial intelligence and worker well-being emphasize that employers should use AI in ways that support transparency, worker well-being, and responsible deployment. For small businesses, that means employees should know what is being monitored, why it is being monitored, and how the information may be used.

AI, Confidential Information, and Data Security

One of the most immediate risks for small businesses is employees entering confidential information into public AI tools. Employees may paste customer lists, payroll data, employee medical notes, disciplinary records, contracts, trade secrets, or internal complaints into AI systems without understanding where that data goes.

This is a serious HR and business risk. Employers should adopt a policy prohibiting employees from entering confidential company information, employee records, customer data, medical information, financial data, or proprietary materials into public AI tools unless the company has approved the platform and confirmed appropriate privacy protections.

A good AI policy should explain that convenience does not override confidentiality. If an employee wants to use AI to draft a memo, summarize a complaint, or analyze employee data, they should use approved tools and remove identifying information unless authorized.

What Is an AI Audit?

An AI audit is a structured review of how an organization uses artificial intelligence and whether those uses create legal, ethical, operational, or employee relations risk.

An AI audit does not have to be complicated. For a small business, it can begin with a simple inventory: What AI tools are we using? Who uses them? What decisions do they affect? What data goes into them? Who reviews the outputs? Are employees or applicants told that AI is being used?

The National Institute of Standards and Technology has developed an AI Risk Management Framework to help organizations manage AI risks. NIST’s framework is built around functions such as governing, mapping, measuring, and managing AI risks. While small employers do not need to become AI engineers, the framework reinforces an important business principle: AI risk should be managed intentionally, not casually.

For HR, an AI audit should focus on whether AI tools affect hiring, scheduling, pay, productivity evaluation, discipline, promotion, termination, leave management, or employee monitoring.

What an HR-Focused AI Audit Should Review

An HR-focused AI audit should examine at least five areas.

First, the employer should identify every AI or algorithmic tool used in the workplace. This includes obvious tools like ChatGPT, but also less obvious tools built into recruiting platforms, payroll systems, scheduling software, surveillance platforms, learning systems, and performance dashboards.

Second, the employer should classify the risk level of each tool. A tool that helps draft internal emails is low risk. A tool that screens applicants, ranks employees, or recommends termination is high risk.

Third, the employer should review data inputs. If the tool uses inaccurate, incomplete, biased, or irrelevant data, its outputs may be unreliable. Bad data creates bad decisions.

Fourth, the employer should review human oversight. Every high-risk AI use should have a named person responsible for reviewing outputs before decisions are made.

Fifth, the employer should document the review. If the EEOC, DOL, TWC, or a plaintiff’s attorney later questions the decision, the employer should be able to show that it did not blindly rely on technology.

The Human-in-the-Loop Rule™

The Unit Consulting recommends a simple standard for employers: The Human-in-the-Loop Rule™.

Before using AI in any HR decision, ask four questions.

Can the Decision Be Explained?

If the employer cannot explain why the AI tool reached a recommendation, the employer should not rely on it for a major employment decision. A black-box recommendation may be convenient, but it is not defensible.

For example, if a tool says Candidate A is a “better fit” than Candidate B, the employer must know what that means. Does it mean the candidate has required experience? Does it mean the resume used certain keywords? Does it mean the candidate resembles past hires? If the employer cannot explain the reasoning, the decision is risky.

Could This Disproportionately Affect a Protected Group?

Employers should ask whether the AI tool might screen out people based on race, sex, national origin, religion, disability, age, pregnancy, or other protected characteristics. The tool may not ask for protected information directly, but it may use proxies that create similar results.

For example, an algorithm may favor applicants from certain schools, zip codes, career paths, or work histories. If those factors are not job-related, they may create legal risk.

Has a Human Reviewed the Recommendation?

A trained human should review AI recommendations before they affect hiring, promotion, discipline, termination, or pay. This person should verify the facts, check the policy, consider exceptions, and confirm that the decision is consistent with how similar employees or applicants have been treated.

Human review should be meaningful. It should not be a rubber stamp.

Would You Defend This Decision to the EEOC, DOL, TWC, or a Judge?

This is the practical test. If an employer would feel uncomfortable explaining the AI-assisted decision to a government agency, the decision needs more review.

A defensible decision should be based on facts, policy, job-related criteria, consistency, and documentation. If the only explanation is “the software said so,” the employer is not ready.

Building an AI Workplace Policy

Every employer using AI should have a written AI policy. This does not need to be complicated, but it should clearly explain how employees may and may not use AI at work.

A strong AI workplace policy should address approved tools, prohibited uses, confidentiality, employee data, customer data, human review, accuracy checks, and disciplinary consequences for misuse. It should also explain that AI-generated content must be reviewed before being used in official business communications or employment decisions.

For HR purposes, the policy should clearly prohibit employees from using AI to make final decisions about hiring, discipline, termination, pay, leave, accommodations, or complaints without management or HR review.

AI Policy Language Employers Should Consider

A practical AI policy might state that employees may use approved AI tools to support productivity, drafting, research, and administrative tasks, but may not enter confidential, personal, medical, financial, employee, or customer information into unapproved AI systems. It may also state that AI-generated content must be reviewed for accuracy, professionalism, confidentiality, and compliance before use.

For managers, the policy should state that AI tools may not be used as the sole basis for employment decisions. Any AI-assisted employment decision must be reviewed by an authorized human decision-maker and documented according to company policy.

This type of policy gives employees permission to innovate while protecting the business from careless use.

Questions Employers Should Ask Before Buying AI Software

Before buying an AI hiring, monitoring, scheduling, or performance tool, employers should ask direct questions.

What employment decisions will this tool affect? Has the tool been tested for discrimination or adverse impact? Can the vendor explain how the tool reaches recommendations? What data does the tool collect? Does the tool use employee or applicant data to train future models? Can applicants or employees request accommodations? Does the tool comply with ADA, Title VII, ADEA, and FCRA requirements where applicable? Can the employer export records if a claim is filed?

If the vendor cannot answer these questions clearly, the employer should slow down. A software demo is not a compliance review.

Training Managers on AI Use

AI risk often starts with managers. A manager may use AI to draft a write-up, summarize an employee complaint, screen candidates, or evaluate productivity without understanding the legal consequences.

Employers should train managers on basic AI rules. They should know not to enter confidential employee information into unapproved tools. They should know that AI content may be inaccurate. They should know that AI recommendations cannot replace policy, documentation, or human judgment.

Most importantly, managers should understand that AI does not remove accountability. If a manager uses AI to create a biased, inaccurate, or retaliatory employment decision, the business is still responsible.

AI and Workplace Investigations

AI can be helpful in workplace investigations, but only in limited ways. It can help organize timelines, summarize policies, draft interview questions, or identify inconsistencies in documents. However, employers should be extremely careful about entering witness statements, medical information, harassment allegations, or confidential employee details into public AI tools.

AI should not determine whether harassment occurred, whether an employee is credible, or whether discipline is warranted. Those conclusions require human judgment, factual evaluation, and policy analysis.

In sensitive matters such as harassment, discrimination, retaliation, theft, violence, or safety violations, AI should support the process, not control it.

AI and Performance Reviews

AI may help managers draft performance review language, organize examples, or compare goals against outcomes. However, AI-generated performance reviews can become generic, inaccurate, or unfair if managers do not provide real facts.

A strong performance review should be based on actual job expectations, measurable outcomes, documented examples, and consistent standards. If AI writes a review without accurate input, the result may sound professional but still be useless or misleading.

Employers should require managers to verify AI-assisted review content before delivering it to employees.

AI and Wage and Hour Compliance

AI scheduling and productivity tools can also create wage and hour risk. If a non-exempt employee uses AI tools after hours, answers AI-assisted work messages, reviews AI-generated tasks, or performs remote work through AI platforms, that time may be compensable under the Fair Labor Standards Act.

The U.S. Department of Labor’s resources on wages and the Fair Labor Standards Act make clear that non-exempt employees must be paid for all hours worked. AI does not change that requirement.

Employers using AI-enabled scheduling, timekeeping, or productivity tools should ensure that all work time is recorded and paid. They should also train employees not to perform off-the-clock work using AI tools.

AI and Employee Trust

Responsible AI is not only about legal compliance. It is also about trust. Employees may become concerned when employers use monitoring tools, productivity scores, automated scheduling, or AI-generated performance feedback.

If employees believe AI is being used secretly or unfairly, morale can suffer quickly. Employers should communicate clearly about how AI is used, what it measures, and how humans remain involved.

Transparency does not mean revealing every technical detail. It means employees should not be surprised that AI is being used to evaluate or affect their work.

Future-Proofing Your Business Against AI Liability

AI regulation will continue to develop, but employers should not wait for perfect legal clarity before adopting safeguards. The safest businesses will be those that implement responsible AI practices now.

At minimum, employers should create an AI policy, identify existing AI tools, restrict confidential data use, train managers, require human review, assess bias risk, and document high-risk AI-assisted decisions.

The goal is not to avoid AI entirely. That would be unrealistic and unnecessary. The goal is to use AI in a way that improves productivity without weakening compliance, fairness, or employee trust.

Practical Employer Checklist for Responsible AI Use

Before using AI in the workplace, Texas employers should be able to answer the following questions:

• What AI tools are we using?
• What employment decisions do they affect?
• Are employees or applicants being screened, scored, ranked, monitored, or evaluated?
• Has the tool been reviewed for discrimination risk?
• Does the tool affect individuals with disabilities, older workers, or other protected groups?
• Is confidential employee or customer information being entered into AI systems?
• Is a human reviewing all AI-assisted employment decisions?
• Do we have a written AI policy?
• Have managers been trained?
• Would we be comfortable explaining this process to a government agency?

If the answer to several of these questions is “no,” the business is not ready to use AI in high-risk HR decisions.

The Bottom Line for Texas Employers

AI can help employers work faster, organize information, improve communication, and identify patterns. But AI also creates serious HR risk when it is used without oversight.

Texas employers should remember the core principle: AI should assist human decision-making, not replace it. The employer remains responsible for the outcome, even when a vendor built the tool, a system generated the score, or a manager relied on a recommendation.

The businesses that benefit most from AI will not be the ones that use it the most aggressively. They will be the ones that use it responsibly, transparently, and with strong human oversight.

How The Texas HR Experts at The Unit Consulting Can Help

At The Unit Consulting, we help Texas employers modernize their HR practices without creating unnecessary compliance risk. Artificial intelligence can be a powerful tool, but only when it is supported by strong policies, trained managers, responsible documentation, and human decision-making.

We can help your business create an AI workplace policy, review AI tools for HR risk, train managers on responsible AI use, update hiring and discipline procedures, and build an AI audit process that protects your business as technology continues to evolve.

If your company is already using AI or considering AI tools for hiring, monitoring, performance management, or employee documentation, now is the time to build the right guardrails.

The Unit Consulting helps Texas businesses use modern tools without losing control of the human side of HR.