Berlin (dpa) – For decades, applying for jobs usually ran to a set formula: Gather your certificates, update your CV, craft a tight but stylish cover letter and fire it all off to the prospective employer.
Today, it still goes more or less like this – but before a superior sees the application, you may first have to win over a bot. Artificial intelligence (AI) is rapidly gaining ground in this area of our working lives as well.
«Theoretically, AI can carry out the entire recruiting process on its own,» says Annika von Mutius, founder of the Berlin-based HR tech start-up Empion, which offers AI-based solutions.
«AI finds out what the ideal candidate for the job looks like, how potential applicants can be reached and how they want to be approached.»
AI can also handle communication. «A robot is good at answering typical questions from applicants,» says Kai Helfritz from the German Society for Human Resource Management (DGFP).
For example, how many days of holiday does the company offer, and does it help with transport costs? An AI chatbot has the answers.
Once the application deadline has passed, AI filters the best applicants from the virtual pile. It could even conduct interviews, analyze applicants’ facial expressions and gestures and then draw conclusions – theoretically, at least. We are not quite there yet, say the experts.
AI is currently only used in some areas of recruiting, but has proven particularly effective in sourcing and screening.
«Sourcing means: Whom am I targeting with my job advertisement and how do I address them? Screening is used in the next step,» explains Mutius.
Once the applications are in it’s all about selection. Most of the time, AI is responsible for a pre-selection, the results of which then land on the recruiter’s desk.
So how can applicants stand out from the crowd under AI scrutiny?
Professor Enzo Weber of Germany’s Institute for Employment Research (IAB) advises including keywords from the job ad in your cover letter: «Formulate precisely, actively and back up your statements with concrete examples.»
One time-honoured truth means applicants won’t go far wrong with this approach: «Getting straight to the point and not formulating too playfully is a good idea in job applications anyway,» says Weber. «The amount of relevant information in relation to the overall text should be as high as possible.»
Helfritz of the DGFP advises: «Write in simple sentences, structure the important information well. It may even be useful to use bullet points.»
Mutius also recommends not to focus too much on tasks mastered when describing your abilities, but on what you learned in the process. She even advises sharing personal information, such as values.
The guidelines all make sense, but only AI itself knows what it likes. When asked what a good application should look like, ChatGPT endorsed the answers given by the experts, while adding some tips: Avoid pictures and graphics as they can be misinterpreted, and use common formats like Word or PDF for easy processing.
ChatGPT also notes that AI systems don’t only search for keywords, but also for patterns and context. So make your application appealing to both AI systems and human readers.
It’s all very well to anticipate the system’s wants, but what exactly is such data-driven personnel decision-making doing to our working world? Is discrimination on the rise, or maybe decreasing?
Theoretically, AI may discriminate more, less or the same as a human being, according to Mutius. On the positive side, an algorithm has no prejudices. But AI systems are trained with real data from the past and put characteristics into context, which can disadvantage applicants because of age or gender, for example.
«The AI architecture as a basis for data-driven decision-making must allow for changes in society,» says Mutius. In other words, conclusions drawn by AI that lead to discrimination cannot be the basis for decisions, and the architects of AI must ensure this.
This renders a human touch behind the architecture crucial to success. «Humans still have to do a quality check to prevent any discrimination,» says Helfritz from the DGFP.
One thing is certain: Applicants need not fear AI in the HR sphere. A good, clearly structured application is also a good application for AI. And once you reach the person-to-person talking stage, it’s back to what it’s always been about: Is there a human fit between us? Is the chemistry right?
If you want to play safe until then, you can beat the computer at its own game. After all, AI doesn’t just work for companies – it also writes perfect, AI-optimised job applications. Only the facts, as always, must be verified by humans.
By Marie von der Tann, dpa