AI in the Workplace: A Practical Guide for Employers

Artificial intelligence is becoming an increasingly important issue for employers. Many businesses are already using AI, or considering using AI, in areas such as recruitment, HR administration, document production, customer service, performance management, workplace monitoring, data analysis and internal communications. 

For employers, AI is not simply an IT or productivity issue. It also involves understanding how AI is being used across the organisation, what information is being entered into AI tools, and whether existing workplace policies are fit for purpose. AI can create legal, commercial and management risks if handled incorrectly. However, when approached properly, it can also help employers improve efficiency, reduce administrative burden and support better decision-making. 

This article considers the key issues employers should be aware of when using AI in the workplace. 

What Is AI in the Workplace? 

AI in the workplace can cover a wide range of tools and systems. It may include generative AI tools used to draft emails, policies, meeting notes or reports. It may also include recruitment software, automated scheduling tools, chatbots, transcription software, productivity monitoring systems, HR analytics platforms and AI tools built into existing workplace software. 

Some AI use will be obvious. For example, an employer may decide to introduce a particular AI platform for document drafting or customer service. Other AI use may be less obvious, particularly where employees are using free or publicly available AI tools informally to help with day-to-day work. 

The important point for employers is that AI may already be in use within the business, even where no formal decision has been taken to introduce it. 

Employers should therefore avoid assumptions and instead focus on understanding how AI is actually being used, who is using it, what information is being processed and whether proper safeguards are in place. 

Why Should Employers Take AI Seriously? 

There are several reasons employers should take AI seriously. 

First, there is the legal risk. If AI is used in recruitment, performance management, redundancy selection, disciplinary issues or workplace monitoring, it may create risks relating to discrimination, data protection, privacy, unfair dismissal and employee relations. The UK does not currently have one single AI-specific law covering AI as a technology; instead, AI is regulated through the context in which it is used and through existing legal frameworks.  

Secondly, there is a management risk. AI-generated information may be inaccurate, incomplete or misleading. If managers rely too heavily on AI without applying proper judgment, this may lead to poor decisions and avoidable disputes. 

Thirdly, there is a commercial opportunity. Used properly, AI can help employers save time, improve consistency and reduce administrative pressure, transforming the workplace. Acas has reported that one third of employers think AI will increase productivity, but has also emphasised that employers should develop clear workplace policies and consult workers and representatives when AI is introduced.  

The Starting Point: Understanding How AI Is Already Being Used 

Before introducing an AI policy, employers should first understand how AI is already being used in practice. 

In many workplaces, employees may already be using AI informally. This could include using AI to draft emails, summarise documents, prepare reports, generate marketing content, produce meeting notes, analyse spreadsheets, answer customer queries or help with HR documents. 

This may be happening without senior management, HR or IT being fully aware of it. 

That creates a problem. An employer cannot properly manage the risks of AI if it does not know: 

• which AI tools are being used;  

• who is using them;  

• what information is being uploaded;  

• whether personal data or confidential information is being processed;  

• whether AI is being used in recruitment or employee management;  

• whether managers understand the risks and benefits of AI adoption;  

• whether employees have received guidance or training;  

• whether existing policies deal with AI.  

For that reason, a practical first step is to carry out an internal AI audit or staff questionnaire. 

Free AI Workplace Questionnaire for Employers 

To assist employers, we have prepared a free AI Workplace Questionnaire which can be used as a starting point for assessing how AI is currently being used within an organisation. 

The questionnaire is designed to help employers identify: 

• whether employees are already using AI tools;  

• which AI platforms are being used;  

• what tasks AI is being used for;  

• whether confidential information, personal data or client information is being entered into AI tools;  

• whether AI is being used for HR, recruitment or employee management purposes;  

• whether employees understand the risks of AI-generated content;  

• whether managers need training;  

• whether an AI policy is required;  

• whether existing data protection, confidentiality, IT and disciplinary policies need to be updated.  

The questionnaire can help employers move from a general concern about AI to a more informed and practical plan. It can also assist with preparing an AI policy that reflects how the business actually operates, rather than relying on a generic policy that may not address the real risks. 

A link to the free AI Workplace Questionnaire can be found here: [insert link]. 

AI and Recruitment 

Recruitment is one of the areas where AI can appear particularly attractive. AI tools may be used to draft job adverts, screen CVs, rank candidates, assess tests, analyse interview responses or communicate with applicants. 

However, recruitment is also one of the highest-risk areas. 

The Information Commissioner’s Office has recently highlighted the use of AI and automation in recruitment, including tools used to sift CVs and score online assessments. The ICO has also emphasised the importance of transparency and the need for employers to use people’s personal information fairly when automated tools are used in recruitment.  

Employers should be careful not to assume that an AI recruitment tool is fair simply because it appears objective. AI systems may reflect bias in the data they have been trained on, the criteria they apply or the way the results are interpreted. 

Before using AI in recruitment, employers should consider: 

• what the tool is actually doing;  

• what data it uses;  

• whether candidates are told that AI is being used;  

• whether the tool has been tested for bias;  

• whether reasonable adjustments can be made for disabled candidates;  

• whether a human manager reviews the outcome;  

• whether rejected candidates can challenge or query the decision;  

• whether the employer can explain how the decision was reached.  

AI may assist recruitment, but it should not become an unexplained barrier between the candidate and the job. 

AI and Discrimination Risk 

One of the most significant employment law risks is discrimination. 

AI tools may appear neutral but still produce discriminatory outcomes. For example, an AI recruitment tool may disadvantage candidates with career gaps. This could affect women who have taken maternity leave, employees with disability-related absence, or candidates with caring responsibilities. 

Similarly, a performance tool may favour employees who work longer visible hours, which may disadvantage disabled employees or employees with childcare responsibilities. A communication analysis tool may also misinterpret different communication styles, including those linked to neurodiversity. 

Employers should therefore consider whether AI tools could create risks of direct discrimination, indirect discrimination, discrimination arising from disability, failure to make reasonable adjustments or harassment. 

This is particularly important where AI is used in recruitment, promotion, redundancy selection, performance management, disciplinary processes or dismissal decisions. 

The safest approach is to ensure that AI informs decisions rather than makes them, and that managers are trained to question AI-generated recommendations rather than simply accept them. 

AI, Confidentiality and Data Protection 

One of the most immediate risks for employers is employees using AI tools without understanding confidentiality and data protection implications. 

For example, an employee might upload client information, employee data, financial information, internal strategy documents or commercially sensitive material into a public AI tool to summarise or rewrite it. That may create serious confidentiality, data protection and commercial risks. 

Employers should make clear what employees can and cannot input into AI tools. This should include rules on: 

• personal data;  

• special category data;  

• client or customer confidential information;  

• commercially sensitive information;  

• privileged material;  

• passwords and security information;  

• employee relations documents;  

• disciplinary, grievance or sickness absence material;  

• contracts and legal documents.  

Employers should also decide whether employees are permitted to use public AI tools, approved internal tools only, or no AI tools unless authorised. 

A short AI policy is now becoming a practical necessity for many employers. However, before preparing that policy, employers should ideally gather information about how AI is already being used across the business. 

AI and Workplace Monitoring 

Employers may be tempted to use AI to monitor productivity, attendance, location, keyboard activity, emails, calls, customer interactions or employee behaviour. This may be especially relevant where employees work remotely or where employers want greater visibility over performance. 

However, workplace monitoring can be legally and culturally sensitive. 

Employers should be clear about: 

• why monitoring is necessary;  

• what is being monitored;  

• whether there is a less intrusive alternative;  

• whether employees have been told;  

• how long the data will be kept;  

• who will have access to it;  

• whether the monitoring is proportionate;  

• whether a data protection impact assessment is needed.  

Excessive monitoring can damage trust, morale and retention. It can also increase the risk of grievances, data protection complaints and employment tribunal claims. 

Employers should avoid introducing monitoring technology simply because it is available. The question should be whether it is necessary, proportionate and properly explained. 

Questions relatives aux performances, à la conduite et aux capacités 

One of the most difficult areas for employers is using AI in performance management, conduct or capability issues. 

Employers are not prevented from using data or technology to support management decisions. However, they should take care before relying on AI-generated information as the basis for formal action. 

Performance management still requires context. A drop in productivity may be caused by workload, unclear instructions, disability, poor management, workplace conflict, caring responsibilities, system issues or other legitimate factors. 

Before using AI-generated information in a disciplinary, capability or redundancy process, employers should consider: 

• whether the data is accurate;  

• whether the employee knows the data is being collected;  

• whether the employee has had an opportunity to respond;  

• whether there may be an explanation for the data;  

• whether a manager has properly reviewed the position;  

• whether there is a risk of discrimination;  

• whether reasonable adjustments should be considered;  

• whether the decision can be justified without relying solely on AI.  

If employers move too quickly to formal action without considering these issues, they may expose themselves to avoidable legal risk. 

Introducing AI and Employee Relations 

Introducing AI can create anxiety among employees. Staff may worry that AI will replace jobs, increase monitoring, reduce human judgment or change the nature of their role. 

Employers should therefore treat AI implementation as an employee relations issue, not just an IT project. 

In some cases, introducing AI may affect job duties, performance expectations, reporting lines, working methods or staffing levels. If employees are expected to use AI as part of their role, that may require training and may, depending on the circumstances, involve changes to terms and conditions. 

Employers should consider: 

• whether employees need to be consulted;  

• whether contractual changes are required;  

• whether job descriptions need updating;  

• whether training is needed;  

• whether AI affects performance expectations;  

• whether there is any redundancy risk;  

• whether managers can explain the purpose of the change.  

A transparent approach is likely to reduce resistance and improve adoption. 

Creating an AI Policy 

An AI policy does not need to be overly complicated. It should be practical, clear and tailored to how the organisation actually works. 

An effective AI policy should explain: 

• which AI tools are approved;  

• when employees may use AI;  

• when employees must not use AI;  

• what information must not be entered into AI tools;  

• the need to check AI-generated content for accuracy;  

• when human review is required;  

• how AI may be used in recruitment or HR processes;  

• how AI may be used in performance management;  

• whether AI monitoring is permitted;  

• how employees should report concerns;  

• the consequences of misuse.  

The policy should sit alongside existing policies on IT, data protection, confidentiality, disciplinary procedures, recruitment and social media. 

The best AI policies are not generic. They should be based on the employer’s actual use of AI, the type of data handled by the business, the level of employee access to AI tools, the sector in which the employer operates and the risks identified through any internal audit or questionnaire. 

Les erreurs courantes des employeurs 

Les employeurs doivent être particulièrement attentifs à éviter les erreurs suivantes : 

• introducing AI without telling employees;  

• allowing employees to use public AI tools without rules;  

• assuming that AI-generated content is accurate;  

• relying on AI to make recruitment decisions without human review;  

• using AI performance data without considering context;  

• failing to check for discrimination risk;  

• allowing confidential information or personal data to be entered into AI tools;  

• failing to update data protection, confidentiality and IT policies;  

• using AI monitoring tools without assessing whether monitoring is proportionate;  

• failing to train managers;  

• failing to consult where AI changes roles, expectations or working practices;  

• introducing a generic AI policy without first understanding how AI is actually being used.  

Many disputes arise not because the employer intended to misuse AI, but because there were no clear rules, no proper oversight and no one had paused to consider the risks before the technology was used. 

Mesures pratiques pour les employeurs 

Employers considering AI in the workplace should consider the following steps: 

1. Audit where AI is already being used in the business.  

2. Use an AI workplace questionnaire to understand employee usage, risk areas and training needs.  

3. Identify what employee, candidate, client or customer data is being processed.  

4. Decide which AI tools are approved and which are prohibited.  

5. Carry out a risk assessment before using AI in recruitment, monitoring or HR decision-making.  

6. Consider whether a data protection impact assessment is required.  

7. Consult employees where AI affects roles, working practices or monitoring.  

8. Train managers on the limits of AI-generated information.  

9. Keep human oversight in all important employment decisions.  

10. Update employment contracts, policies and staff handbooks where needed.  

11. Review AI use regularly as the technology and legal landscape develops.  

Taking these steps can reduce legal risk and help employers use AI in a controlled and commercially sensible way. 

Conclusion 

AI in the workplace is an issue employers cannot afford to ignore. Handled poorly, it can lead to confidentiality breaches, data protection complaints, grievances, discrimination claims, unfair dismissal risks and reputational damage. Handled properly, it can help employers improve efficiency, support managers and modernise working practices. 

The key for employers is to act early, avoid assumptions and understand how AI is actually being used within the organisation. AI does not remove the need for human judgment, fair procedures or clear workplace policies. 

At Ronald Fletcher Baker, our Employment Team advises employers on workplace policies, data protection issues, recruitment, disciplinary and grievance procedures, redundancy, discrimination risk and employment tribunal claims. If you are considering introducing AI into your workplace, or if concerns have already arisen about the use of AI, we can help you assess the risks and identify a practical way forward.