NEW APPROACHES IN HRM

59 experience. These inconsistencies are reflecting the generational differences of actors involved. The most positive returns are seen in the domain of employer branding, giving the company a label of being up to date, progressive and innovative. On the other side, some of the applicants had less positive perception, believing that if company has no time to contact them in person, it means that for them he is not important. From the Trisma`s point of view, they concluded that when incorporating a new digital practice in HR processes, the starting point should not be from the most complex technology available but rather going one step at the time. 2. 3. SIDE EFFECTS OF DIGITAL HR – POSSIBLE IMPROVEMENTS Promoters of chat bots claim that the biggest benefit of using them in the process of selection is the objectivity, consistency and transparency. It definitely cut down a workload from the humans in the process of recruiting. Being able to process multiple candidates’ simultaneously and continuously, it reduces time and save money. So, „ recruiting chat bots are becoming an integral part of the talent acquisition process ” (Ghosh, 2019). Although it is expected that chat bots are impartial and can reduce the prejudices that humans ‘are prone to, it is not always a reality. The famous case of Amazon`s biased recruitment bot , shows that it is possible. This case teaches us that technology can do only what we programed it to do. The Amazon`s system was not free from gender bias that was incorporated into an algorithm based on the learned pattern from the data obtained from the previous selections. That is why Biswas (2019) said that AI should help HR decision making faster and not instead of it while it is much easier to exclude bias from the algorithm than from the mind of the person. The problem of decision making is also present in the people analytics. Although it is helpful having available the great amount of complex data processed and prepared for evaluation, the problem lies in the quality and validity of data and in consistency of metrics and its compatibility at different platforms, while protecting data privacy . Hans (2018) see the problem of synchronization of the huge amount of data generated in different systems. Orlikowski and Scott (2008) imply that technology open two issues. First, there is a question of system`s infrastructure