.svg)
.jpg)
Digital voice assistants, personalised marketing, chatbots: even though artificial intelligence (AI) still has to contend with a negative image, it has already become part of our everyday lives in some areas. Even in the detailed processing of large amounts of data, i.e. in big data analytics, learning algorithms offer promising opportunities to speed up processes without having to reprogram every time conditions change.
This can also be helpful in recruiting: for example, in the search for suitable candidates and the processing and assessment of applications. But for which (sub-)processes of recruitment should AI be used? And for which ones should we rather not? I would like to find an answer to these questions today.
What exactly is Artificial Intelligence?
The term AI is now used in many areas - sometimes with very different interpretations. This is because artificial intelligence is an entire subject area, namely a branch of computer science. That's why the understanding of it varies so much. In classic science fiction films, the so-called strong AI often appears: Machines that learn on their own and can solve any kind of problem. However, this still remains pure fiction.
In reality, only weak AI has been used so far: complex algorithms that are programmed to solve certain questions and problems based on learned contexts. They are already appearing in everyday life in mobile phones and computers whenever rule-based programmes are not sufficient for the complexity of a problem - for example in automatic face recognition. The more input the algorithm gets, the more accurately it can recognise individual faces and assign people.
Once the AI has been programmed to apply one or more basic rules, it can internalise them more and more as the amount of data grows and gain in accuracy. This principle offers enormous potential for the economy. The German government estimates that AI will account for more than a third of future value creation by 2025. And in recruiting, too, people are slowly becoming aware of the possibilities of this technology.
Possible Applications in Recruiting
Due to the automated processing of large amounts of data, basic selection decisions can be made much faster by artificial intelligence than by "real" recruiters. For example, career networks can be scanned by corresponding algorithms to find suitable candidates. Through the subsequent feedback from HR managers on the suitability of the candidate selection, AI can optimise its decision-making process further and further.
Artificial intelligence can also support the application process: Applicants can turn to chatbots for simple and general questions and in some cases even make appointments with them. The simplicity and speed of this measure can significantly improve the candidate experience.
In so-called CV parsing, AI takes over the initial review of the application documents received. If you give it a few criteria for classification, it can support the pre-selection of candidates by assessing the quality of letters and documents and creating suggestion lists by means of matching.
Artificial intelligence can help with the development and selection of questions in preparation for job interviews. This not only saves time, but also has the advantage that different applicants can be better compared with each other.
So far, the use of AI in job interviews has been rare and extremely controversial. It is not only possible to use a video chatbot as a communication medium. Some (so far few) employers also use the technology "covertly" to precisely analyse speech, intonation and facial expressions in order to draw conclusions about character traits and attitudes.
The Opportunities
There is no question: parts of the recruitment process can be accelerated through the use of AI. Especially when searching for suitable talent for a difficult-to-fill position, significantly more people can be considered overall. The higher quantity of potential candidates presumably also increases the quality of the talent ultimately hired. In addition, the selection decisions are precisely comprehensible and consistent in themselves, as they are based on concretely defined criteria that are simply reproduced by the AI.
Artificial intelligence can therefore take over the otherwise very time-consuming tasks in recruiting from human recruiters. This reduces the workload and ultimately leaves more time for the personal support of applicants. Of course, this also has advantages for the applicants, because the company's feedback times can be drastically reduced.
The Risks
However, the question of "better" selection decisions through AI must be considered from two sides. On the one hand, it can be argued that algorithms make fairer decisions because they are free of human prejudice, unconscious bias, emotion and arbitrariness - and thus do not discriminate. However, AIs learn through injected data sets: Due to biased training data or prejudices of the developers, the AI could possibly draw false connections, make discriminatory decisions itself and scale up these errors due to the large amount of processed data.
The reproduction of always the same selection criteria can also lead to a decrease in diversity in the company, which is actually an important enrichment for the company culture. So here, too, programming and training data are very important - if mistakes or imbalances occur, the consequences can be severe.
The field still lacks transparency and ethical guidelines that set a framework for the use of AI in recruiting that is acceptable to all. I see the use of AI in job interviews as particularly critical: I am not only concerned with the question of ethics and morals, but also with the fact that interpersonal skills such as empathy, sympathy and the communication of respect and appreciation are particularly important in direct contact with candidates. If it is not possible to respond individually to candidates and show commitment, interest in the vacant position will quickly be lost.
The Data Protection Issue
In Germany, the protection of personal data plays a major role. This also applies to recruiting, which is primarily about the evaluation and interpretation of such personal data. Therefore, strict regulations apply that severely restrict the use of AI in recruiting without the candidate's consent. Applicants must agree to the use of the data, including for use as training data. It is also important that such consent is voluntary - and let's face it, if the talent is excluded from the application process in the event of a rejection because their data cannot be processed, then there can be no question of a voluntary decision.
Therefore, the use of AI in recruiting will only work if it is completely transparent, the work of AI is permanently monitored and adjusted in case of weaknesses or errors. In addition to solving data protection problems, this is likely to bring another advantage: Because presumably this will also increase acceptance of AI in recruiting.
A Vision of the Future: AI suggests, People decide
The goal of recruiting should always be a "perfect fit" between people and companies. Of course, this includes hard skills that are needed by the company and are indispensable. But it is equally important that employees have the desired soft skills, that their character fits the company, that common values are present and that new team members can identify with their work.
When it comes to the first part, i.e. the "hard facts", well-programmed and trained AIs can certainly be a helpful support. But for the assessment of soft skills and character, real people are needed who quickly develop a feeling for their counterparts, who are empathetic and can interpret interpersonal signals. Not least because they themselves create human closeness with which applicants can identify.
That's why I think it makes sense and holds promise for the future that AI is used in certain areas of recruiting: If possible, where it is about the transfer and evaluation of data and facts and not about their interpretation. This saves time for recruiters. Time that they can use for interpersonal contact, for example to exchange information with their candidates, to conduct interviews or to organise onboarding.
Artificial intelligence can support by sorting, calculating probabilities, making suggestions based on data sets - but the final hiring or rejection decision should always be made by humans. So the future lies in optimised collaboration between people and technology, leveraging strengths and balancing weaknesses.
Please also read the article by our partner alphacoders: "ChatGPT: use cases for tech recruiting".
