Recruitment by Algorithm: Are We Trusting Code with Care?
We’re inviting algorithms into recruitment, hoping they’ll help sort candidates faster, reduce admin, and make hiring more data‑driven. But in healthcare, where values and judgment matter as much as credentials, handing over decisions to machines can be risky.
Real Risks Behind the Tools
In 2024–25, there are still 111,000 job vacancies in adult social care, with an estimated 7 % vacancy rate in independent providers. (BCOP, based on Skills for Care data) [1] That’s a huge stress on staffing capacity. Yet, many organisations are turning to AI screening to help fill roles quickly.
But the problem is that AI models reflect the biases in their training data. For instance, a University of Washington study showed that large language models used in hiring ranked resumes differently based on perceived race or gender from names alone. [2] That matters in healthcare, bias in early screening can exclude capable caregivers before they even get a chance.
A review in Nature Digital Medicine also warns of bias slipping in through every stage, from feature selection to deployment. [3] And when algorithms start deciding who’s “fit,” we risk losing the human judgment that catches what code can’t see.
The Promise If We Build It Right
AI in hiring doesn’t need to be a villain. In fact, used well, it can augment fairness, speed, and insight. A recent article in the World Economic Forum argues that AI can make recruitment more strategic and inclusive, if humans lead, not follow. [4]
We just must put the right guardrails in place:
- Algorithms should highlight, not decide.
- Models must be explainable and transparent.
- Continuous feedback loops are essential, monitor which hires succeed, which don’t, and adjust.
- Ensure data diversity and inclusive model design so the algorithm doesn’t just mirror systemic bias.
- Preserve space for human interview and judgment to catch nuance.
In the end, culture, trust, and relational care can’t be automated. Algorithms can help us filter, flag, support. But they can’t replace the conversation, empathy, and shared values that make care human.
So here’s what I keep asking:
If we trust code to screen caretakers, who’s preserving character, judgment, belonging?
If your organisation is ready to explore AI‑driven recruitment, without losing the human part, I’m here to help. Email me at [email protected].
References
- BCOP / Skills for Care (2025). “111,000 job vacancies … 7% vacancy rate in adult social care.” (based on Skills for Care data)
- University of Washington (2024). “AI tools show bias in ranking job applicants’ names by perceived race and gender.”
- “Bias recognition and mitigation strategies in artificial intelligence.” Nature Digital Medicine (2025).
- “Hiring with AI doesn’t have to be so inhumane. Here’s how.” World Economic Forum (2025).