�ݺ�ߣ

Editor's Notes

• #2: Photo by Ergoneon on pixabay
• #3: Photo by Paul Mawson auf Pixabay  We all love HR Tech. I love HR Tech. I‘m serious! What we‘ve seen coming to the market in the past years is fantastic:  Much has been said about robots in recruiting: Chatbots as the answer to a better candiate experience Matching algorithms replacing traditional job ads Artificial intelligence as the antagonist of human bias But as much as I believe that HR Tech will improve all our lives - as recruiter, as hiring manager as candidate – there are some areas, where machines fall short and nothing will replace us as humans. And this is what I will try to shed some light on. 
• #4: Photo by Rock'n Roll Monkey on Unsplash Let‘s start by taking a look at what machines are particular good at. If you ask me there are two main areas: AUTOMATION DISCOVERY of pattern in data
• #5:  Photo by ainthome on Pixabay  AUTOMATION can be applied in almost all areas of TA funnel. When does Automation work best? When the machine is measuring... ...yes/no, right/wrong answers ...clicks vs no clicks ...application vs no applications ...similarities in % against given profiles It will start a set of predefined set of actions based on the results of above mentioned questions. Machines are here clearly superior to humans as they are able to process MORE information IN PARALLEL, FASTER than any human mind. Machines=1, humans=nil
• #6: Photo by Markus Spiske on Unsplash DISCOVERY Machines can also analyse more information than any human. They don‘t overlook information, they don‘t forget to include data, they pay attention all the time. This enables them to analyse information way better than us. As a result they can detect pattern in data much better than us. Now here‘s the thing: they are better in detecting pattern, but are they really better in INTERPRETING the pattern? We have endless examples of significant statistical relations where human intelligence clearly tells us there is no connection. Think! The relation between suicides and investments in science & space technology in the US.
• #7: What we see here is a seemingly clear relation between US spendings in science, space and technology (red line) and Suicides by hanging, strangulation and suffocation (black line). For the machine it displays a statistically highly relevant correlation where the human mind immediately rejects the correlation as highly unlikely. 
• #8: Photo by Markus Spiske on Unsplash Why is that? Because the machine did/could not consider all relevant information do invalidate the decision, where the human mind already jumps to the conclusion that a correlation between two items is most unlikely and can therefore be discarded. The human mind can draw conclusions where the machines needs information to invalidate. For the foreseeable future the machine will not cover all the information the human mind has and will still need human intelligence to validate discovered pattern. So far so good: machines=2, human=1.
• #9: Let‘s move on to another factor why I think we won‘t loose our jobs as recruiters - and that has to do with the core of our jobs: finding the right people to secure the success of our organizations.
• #10: Success has become increasingly futile over the past years, economic growth is not as predictable as it used to be. There is a term for this world we are living in: VUCA. It stands for VOLATILE, UNCERTAIN, COMPLEX, AMBIGIOUS. Past models and behaviors don‘t necessarily work anymore. Organizations have to increasingly navigate through unknown territory and make decisions either in the absence of sufficient or in the face of overwhelming information. Which is why we need to hire people that have the capabilities to navigate in such an environment.
• #11: And while I love HR Tech - the ability of machines in this context is limited for two reasons: 1) They are looking at the wrong things 2) They are not good at looking at the right things
• #12: Sergio Leone [Public domain], https://commons.wikimedia.org/wiki/File:Clint_Eastwood1.png WRONG THINGS Machines are trained looking back. We feed them with past behaviour that led to success. That is at most a predictor for very short term success under the same conditions. Thus it will not help you finding that future talent your CEO is asking you for.
• #13: Photo by Xuan Nguyen on Unsplash Machines are (equally) biased - because their data input remains to be. Most data used to train algorithms is based on human assessment of success. I don‘t think it need to say any more, everybody think of performance management systems they know and managers feeding these systems for the past decades. You‘ve all heard the Amazon example, I don‘t think I need to go there. So here we see a tie between men and machine: Men is biased and hence is the machine. But it adds to the fact why machines will not be able to asses by themselves until all relevant data is collected objectively enough to come to an valid foundation. When that tipping point is going to be reached remains in the vast mist of the future (someone hand me a crystal ball...) So in this assessment scenario we see a major conflict in the two areas where Machines are naturally strong: Matching (which is AUTOMATION) - doesn‘t work because it‘s looking at the past, Discovery - or more precise Predictive - is weak because the data is flawed.
• #14: Photo by davide ragusa on Unsplash NOT THE RIGHT THINGS Now if it’s not past behaviour that is going to help us - what will???
• #15: Photo by Fares Hamouche on Unsplash Looking at current research there are more or less four factors that qualify as VUCA- or „digital readiness“: # Smart # Learning ability # Self-reflection # Adaptable As we are in the TA profession, how can we assess these - and can they be automated?
• #16: Photo by Fares Hamouche on Unsplash # Smart ✅  # Learning ability ❓ # Self-reflection❓ # Adaptable ❓ How smart a person is can be assessed with proven cognitive tests. These are scientifically validated tests that thanks to HR Tech have become much more candidate friendly than the old paper/pencil tests of the past. Whether one should choose gamified or game based assessments is topic for another talk. At the end these tests are similar to very elaborate yes/no answers with an ability to adapt to the users performance thus they more or less fall into one of the strength areas of machines, automation. The last three are soft skills. Even in given scientific personality assessments the reliability of the findings require a multi dimensional approach to reach a sufficient level for making a hiring decision. No sane diagnostic expert would hire a person only based on their assessment results - they would always amend them with other instruments, most likely interviews. Again this can not be substituted by a machine because what would they measure the candidate‘s answers against? It takes human intelligence - and emotional intelligence - to amend assessment findings and identify the most suitable candidate - not to speak of assessing their personal fit to the organization and convincing them to join (that‘s a whole other story).
• #17: Photo by Andy Kelly on Unsplash So wrapping it up: Machines for the foreseeable future will help us in most TA fields due to automation, yet the least in assessment. Here we need the human # to train algorithms # to ethically define and decide their training basis # to amend machine findings # to bring emotional intelligence to the recruiting process After all joining an organizations is much more than the sum of rational 0I decisions but very much an emotional process on both sides! Think of HR Tech not as the autonomous driving vehicle – it remains your driver assistant system! Familiarize yourself with it, let it support you – but keep your hands close to the steering wheel.
• #18: Or as Elbert Hubbard, American writer and philosopher put it at the beginning of last century: "One machine can do the work of fifty ordinary men. No machine can do the work of one extraordinary man."
• #19: Bild von Gerd Altmann auf Pixabay