The AI landscape in HR is evolving rapidly. Once dominated by questions like "should we use AI?" we are seeing the conversation mature to "how do we use it safely and responsibly?" This shift was the foundation of a recent panel hosted by (us) Warden AI, featuring leading voices from HR practitioners and HR Tech innovators.
From Hype to Helpful: Where AI is Making a Difference
Sultan Saidov, Co-Founder and President of AI talent platform, Beamery, kicked off the discussion with a high-level overview of how AI is driving real value in HR.
He categorized use cases into two buckets: AI that augments or automates existing tasks (like job description writing or candidate screening), and AI that provides new insights that didn’t exist before.
Crucially, Sultan emphasized that the value of AI isn’t just about efficiency, but rather it’s about empowering HR to be stewards of broader workforce transformation.
For instance, AI may remove repetitive tasks, but HR professionals must then ask: what do we do with the time and talent we've freed up? This will make, Sultan says, more “diamond shaped” organizations vs. pyramid ones.
Dr. Cari Miller from the Center for Inclusive Change echoed this sentiment. She outlined three layers of AI adoption in HR:
- Low-hanging fruit: using tools like ChatGPT to draft policies or compare regulations.
- Pilots: experimenting with vendor tools, often dazzled by slick marketing.
- Deliberate, strategic implementations: identifying a specific problem and using AI to solve it intentionally.
Real Impact or a Mirage?
One question posed to the panel was whether AI is truly displacing workers. While some vendors remark about reducing headcount, the panel struck a more nuanced tone.
Bill O’Connor, Assistant General Counsel at iCIMS, pointed out that automation is helping with routine support tasks and is actually freeing up time without necessarily eliminating jobs.
Sultan added that while automation can lead to staff reductions, many companies are reinvesting in their workforce by retraining talent into higher-value roles, like content labeling or AI oversight. The real takeaway: the impact of AI depends on how consciously an organization manages the transition.
The Risks: Bias, Privacy, and Explainability
With great power comes great responsibility and AI in HR carries significant risk. Bill laid out three major concerns here:
- Bias and Discrimination: Especially in hiring, where regulatory scrutiny is increasing, such as New York City’s Automated Employment Decision Tool (AEDT) law.
- Privacy and Data Protection: Including the source of training data and the handling of sensitive personal information.
- Transparency and Explainability: Organizations must be able to explain how decisions were made, particularly when AI is involved.
Carrie expanded on the risk of vendor literacy, noting that not all AI products are created equal. Some are built on robust, de-biased data sets, while others are cobbled together using freely available models without sufficient safeguards.
For buyers, distinguishing between these requires a healthy degree of skepticism and education.
Navigating Regulation: What HR Needs to Know
With global regulation heating up, organizations must stay alert. The panel highlighted key laws:
- EU AI Act: Classifies HR tech as “high risk,” mandating impact assessments, auditability, and transparency. Read more about how the EU AI Act affects HR Tech platforms.
- NYC Local Law 144: Focuses on bias audits for hiring tools. Here’s everything you need to know about NYC Local Law 144.
- Colorado SB 205: Holds AI accountable for fairness and transparency. Read more about Colorado SB 205 and how it is different from NYC Local Law 144.
- Emerging US State Laws: Including California and Texas, which are introducing AI and biometric-specific employment regulations.
Carrie advised companies to “school up on impact assessments” and start exploring third-party AI bias auditors. As regulations mature, so will expectations for compliance.
Best Practices for Vendors and Deployers
The panel offered guidance on how both AI vendors and HR deployers can implement responsible AI practices:
For Vendors:
- Understand your customers’ compliance obligations.
- Build AI with transparency, 'auditability', and bias mitigation in mind.
- Adopt emerging standards like ISO 42001 and align with frameworks like NIST’s AI Risk Management Framework.
- Consider ongoing audits and publish explainability statements.
- Broadcast your responsible AI efforts: trust marks, certifications, and transparency are key to buyer confidence.
For Deployers:
- Treat AI like any other IT implementation: onboard it, train users, and establish governance protocols.
- Create a cross-functional team to assess tools and track incidents.
- Understand how your chosen vendors are managing AI risk and demand proof.
- Educate staff to avoid AI misuse: especially the accidental sharing of sensitive data into public tools.
- Remember, doing something is better than doing nothing. Start small, iterate, and refine as regulations and your understanding evolve.
Cari Miller delivered a great takeaway on what buyers look for in terms of best practices:
“And so as I look at vendor best practices, I am always looking for the trust marks. I’m always looking for that certification for ISO… that’s huge for me.
But I’m also looking on the website. Tell me what you did. Especially when you have put in this effort for responsible AI. It is not inexpensive. It takes time, it takes money, it takes talent, and it takes knowledge. It’s not easy work.”
Final Thoughts
Each panelist offered practical advice to close the discussion:
- Bill O’Connor, iCIMS: Create a framework aligned with recognized standards. Learn the basics of AI and ask questions.
- Dr. Cari Miller, The Centre for Inclusive Change: Use a cross-functional team, capture every incident (good or bad), and build literacy.
- Sultan Saidov, Beamery: Be transparent. Share what you’re doing and invite feedback to continuously improve your responsible AI program.
As trust in AI grows, especially in controlled enterprise environments, adoption will expand. But success will require more than just implementation. It will require intentionality, accountability, and a culture of continuous learning.
This blog is part of our Warden Webinar Series. We are always on the look out for diverse panelists with a myriad of experience!
To feature as a panelist on our next webinar, send us an email at contact@warden-ai.com. To get access to the full webinar recording, sign up here.
