HR is one of the best places to use AI and one of the riskiest. The best because so much of the job is writing: adverts, offer letters, policies, onboarding packs and answers to the same questions every week. The riskiest because the moment AI starts ranking or rejecting people, you're in the territory regulators watch most closely.

I work with HR teams and business owners worldwide, and I draw a clear line between drafting and deciding. Drafting help comes first. Decision-making tools come later, if at all, and only with checks in place.

HR work AI is good at

These are the tasks I start with, because a person reviews every output before it reaches an employee or candidate.

• Job adverts and descriptions: clear first drafts in inclusive language, checked by the hiring manager.
• Onboarding packs: welcome emails, first-week plans and checklists built from your own documents.
• Policy questions: an internal assistant that answers from your handbook and points people to the right section.
• Interview admin: scheduling emails, question banks and summaries of your own interview notes.
• Reporting: turning HR system exports into a plain monthly summary.

The hiring rules HR teams can't ignore

If you hire in New York City, this one is already law. The NYC Department of Consumer and Worker Protection enforces Local Law 144 of 2021, which bans employers and employment agencies from using an automated employment decision tool unless it's had a bias audit within one year of use, a summary of that audit is public, and candidates or employees have been given notice. Enforcement began on July 5, 2023, and the notice has to be given 10 business days before use.

In Europe, the European Commission lists emotion recognition in workplaces among the practices the AI Act bans, and those bans have applied since 2 February 2025. AI used in employment is also classed as high-risk, with those rules applying from 2 December 2027. So that video interview tool that claims to read a candidate's confidence from their face? In the EU, that's not a feature. It's a prohibited practice.

Where AI shouldn't make the call

Shortlisting, rejections, pay, performance ratings and disciplinary decisions stay with people. AI can summarise information for the person deciding, but it shouldn't decide. Training data carries old biases, and a tool that quietly filters out good candidates does damage you won't see until someone challenges it.

How I bring AI into an HR team

This is the order I follow, so the team gets quick wins without taking on legal risk.

1. We list the writing and admin tasks that take the most time.
2. We agree what AI may and may not do, and write it into your AI policy.
3. I build prompts and templates around your handbook, tone and past documents.
4. Your team uses them for a fortnight and we review the results.
5. Any tool that screens or scores people gets a separate review before anyone buys it.

Choosing HR tools

Plenty of HR software now includes AI features, and some are far better than others. If you're comparing options, my roundup of the best AI tools for HR is a useful starting point. Ask every vendor what their tool decides on its own, what it only suggests, and whether it's been audited for bias.

Questions to ask before you buy an AI hiring tool

Vendors are keen to sell screening, ranking and video interview tools to HR teams. Before you sign anything, ask these questions in writing.

• What does the tool decide on its own, and what does it only suggest to a person?
• Has it had an independent bias audit, and can we see the summary?
• What data was it trained on, and does it use our candidates' data to train?
• Does it analyse faces, voices or emotions? If so, where is that lawful?
• How do we tell candidates we're using it, and how can they ask for a human review?

Keeping employee data safe

HR handles some of the most sensitive data in any business: health notes, grievances, pay and performance. Before anyone pastes that into a chatbot, check whether your tool keeps or trains on what you type, and set a clear rule on what never goes into AI at all. In most teams the safest default is that named employee data stays out of general AI tools unless the business has approved a specific, secured one for that job.

My advice for HR teams using AI

Use AI to write, summarise and organise, and keep people deciding. Start with job adverts, onboarding packs and policy answers, because they save hours and a person checks every one. Before you buy anything that screens, ranks or reads candidates, check the rules where you hire, ask for the bias audit, and tell candidates what you're using.

Working with me

I'm an AI consultant, trainer and speaker with 23 years in marketing. I can set up the workflows, train your HR team, or run a session for managers on using AI fairly. The ways we can work together are on my work with me page.

Frequently asked questions

Who trains HR teams to use AI?

Lilach Bullock, a British AI consultant and trainer with 23 years in marketing, trains HR teams worldwide to use AI for job adverts, onboarding, policy questions and admin.

What can HR use AI for safely?

Drafting job adverts, onboarding packs, policy answers from your handbook, interview scheduling and reports, with a person reviewing every output.

Do AI hiring tools need a bias audit?

In New York City, yes. Local Law 144 of 2021 requires a bias audit within one year of use, a public summary and notice to candidates, enforced since July 5, 2023.

Is emotion recognition allowed in job interviews?

Not in the EU. The European Commission lists emotion recognition in workplaces among the AI Act's banned practices, applied since 2 February 2025.

Should AI shortlist candidates?

I wouldn't let it decide. It can summarise information for the person shortlisting, but rejections and rankings should stay with people.