- Story
“Active sourcing remains human-centered”
14.02.2025 In Switzerland, artificial intelligence (AI) also has a say in who gets hired. We assess this evolution with BFH experts Mascha Kurpicz-Briki and Caroline Straub.
Key points at a glance
- AI-powered recruitment tools are effective but not universally applicable. They are particularly suitable for office jobs.
- The use of AI in HR entails risks and can lead to discrimination.
- Recruitment remains people-centred: AI provides support, but the decisions are made by humans.
How does AI impact HR in Switzerland today?
Straub: Headhunters and HR departments often rely on active sourcing to find potential candidates. According to a study conducted by BFH, all the companies surveyed use active sourcing tools.
However, not many HR specialists use them. Companies generally have restricted licenses for these AI tools and not all recruiters are aware of how to use them optimally. Many still adhere to a traditional approach, i.e. without any AI support.
Our interview partners
Prof. Dr Caroline Straub teaches Leadership and Human Resource Management (HRM) at BFH. She works for the Institute for New Work and specialises in digital HRM, among other things. She is at the head of the research project “Artificial Intelligence applications in HRM: opportunities and risks”
Prof. Dr Mascha Kurpicz-Briki teaches Data Engineering at the Institute for Data Applications and Security IDAS at Bern University of Applied Sciences. Moreover, she is deputy head of the Applied Machine Intelligence Research Group. Her research focusses among other things on fairness and the digitalisation of social and societal challenges. She is at the head of the research project “BIAS: Mitigating Diversity Biases of AI in the Labor Market”
What is active sourcing?
Straub: In active sourcing, companies use online methods to approach and recruit suitable employees.
Digital matching tools (i.e. AI-supported algorithms) are increasingly used to predict the extent to which a job profile matches a potential candidate. Matching tools are used because they can increase process efficiency.
What is particularly important in active sourcing?
Straub: It is essential for sourcers to be familiar with the appropriate keywords (e.g. programming language for IT jobs) to conduct targeted and efficient searches. They must possess a comprehensive understanding of both general keywords and specialised terms used by a group of professionals.
The more they master this, the more successful the search. This is becoming an essential skill for HR specialists. Training in this area is becoming undoubtedly crucial. At the moment, companies don’t often turn to external LinkedIn training courses, for instance. Most of them develop their active sourcing skills through experiential learning or with video tutorials.
What opportunities do active sourcing tools offer recruiters?
Straub: Active sourcing gives them a quick and straightforward overview of potential candidates who can be approached in a next step. They can also find passive candidates who are not actually looking for a new job.
In the first stage, superiors communicate the requirements for the vacant positions. HR then uses active sourcing to provide them with a longlist of suitable candidates. The recommendations are assessed by and with the superiors, and a first (non-automated) contact is made. In other words, active sourcing remains human-centred.
Social stereotypes are also present in language models.
What are the disadvantages of active sourcing tools?
Straub: Our interviews with active sourcers indicate that these tools cannot be used for all jobs, as not all professional groups can be reached on channels such as LinkedIn.
They are primarily suitable for roles in IT, marketing, business development, key account management and engineering. And they are less effective for recruiting individuals who do not work in an office environment, as these professionals typically do not have a LinkedIn account.
Furthermore, highly-sought employees often receive an overwhelming number of requests and may not respond. As a result, recruiters have to reevaluate their processes. For example, alternative campaigns on Facebook, Google Ads, TikTok, or on lorries and restaurant placemats tend to be used to reach professionals in the healthcare sector.
What potential risks should be considered when using active sourcing tools?
Kurpicz-Briki: Bias in AI applications (i.e. distortions or prejudices) can occur due to the underlying selection criteria and data. Researchers at BFH have demonstrated that social stereotypes are also present in language models, which may be used in active sourcing.
This means that simplified filtering and screening methods may result in the overlooking of suitable candidates, while others are wrongly invited to the more cost-intensive selection stages (e.g. interviews). This has financial implications for recruiting companies and entails the risk of discrimination against candidates.
In the Horizon Europe BIAS project, we investigate in particular prejudices (or biases) in recruitment and personnel management applications.
In the event of a discussion with the candidate, personal contact is a priority.
What does that mean concretely?
Kurpicz-Briki: We scientifically test whether a model favours certain profiles and disadvantages others.
Our focus is on technical solutions, but we also try to sensitise those involved – e.g. HR specialists – to the advantages and drawbacks of the AI solutions in place by offering targeted continuing education courses.
Can recruitment with AI-supported tools remain ‘humane’?
Straub: Absolutely. When there is an exchange between recruiter and candidate, the focus is still on personal contact and appreciation.
