What Does Responsible AI Look Like in Talent Strategy?

The regulatory landscape surrounding AI is evolving rapidly, but the technology itself advances even faster. With the introduction of the EU AI Act and hundreds of global regulations, most organisations are only beginning to understand what responsible AI truly requires in a hiring context. The opportunity to build more efficient teams is real, but the associated risks are equally significant. Kevin Morsony, Senior Director of Product Legal at LinkedIn, opened a recent panel by highlighting this exact tension. He sat down with Thom Staight, General Manager of Talent Acquisition EMEA and Asia at Microsoft, and Charlotte Pattison, Global Talent Acquisition Director at Reckitt, to explore the reality of ethical AI adoption.

Watch now

For Pattison, the definition of responsible AI is absolute and non-negotiable. "For me, it has to be fair, it has to be transparent, it has to be ethical," she explained. It is a baseline standard, not a decision that teams can alter on a case-by-case basis. Trust must be deliberately built with both internal business stakeholders and the candidates applying for roles. Staight framed the issue through Microsoft's core operational principle that the business runs on trust. He noted the constant tension between what engineers want to build and what candidates actually feel comfortable using. Risk postures vary widely across different sectors and stakeholder groups. As Staight observed, the ethical red line for some stakeholder groups sits much closer to familiar, older ways of working than either the technology or the regulatory line would suggest. 

This tension becomes tangible when theoretical guidelines meet active hiring workflows. Pattison shared a vivid example of a recent role in a well-known market that generated 4,000 applications. Out of those submissions, over 1,000 appeared to be 100 percent matches because candidates had used AI to edit their CVs to fit the job description precisely. This flooded the system and created a severe challenge for recruiters. It illustrates why responsible AI is a fundamental quality and integrity challenge, not just a compliance checkbox. Microsoft's experience with AI grading tools reveals the same tension from a different angle. Staight shared that Microsoft rigorously tested AI grading tools for inclusivity and the tools passed. The company still chose not to deploy them on active applicants. The reason is instructive: candidate comfort and institutional trust matter just as much as passing a technical test. 

As these tools become embedded in talent processes, organisations must prepare for incoming legal realities. Morsony offered practical guidance rather than a warning. "I think this is going to be ultimately a fairly regulated space. He advised teams to simplify their tech stacks, be rigorous with external vendors, and keep human oversight firmly in the loop. The primary question every organisation should be able to answer is what they will do when a tool they rely on is accused of something that conflicts with their core values.

The questions this panel raised are ones LinkedIn is navigating from the inside. Kevin Morsony builds the legal and ethical frameworks that underpin LinkedIn's AI products — the same verified data foundation and responsible AI principles that underpin everything from Hiring Assistant to Career Hub. The Microsoft partnership adds another layer, given Thom Staight's own account of how seriously Microsoft approaches trust as an operating principle.

Responsible AI in talent strategy is not a compliance exercise that can be completed and filed away. It is an ongoing set of decisions about whose trust you are protecting, which tools you are willing to stand behind, and what you will do when the answers are not straightforward. The organisations that think through those questions now will be the ones best placed to move with confidence when the regulatory and legal environment catches up with the technology. 

Want to explore how LinkedIn can support your team? Get in touch.