Leaders should be careful not

I grew up to follow the tennis legs with my dad, but I only recently returned this sport thanks to another encouraging enthusiastic this game in the family, which is my wife. So it may have been normal to look like me today that the current stars of the game, the incredible ingenuity they enjoy, do not send the ball with the same power as characterized by Beit Sambras or Gouran Ivanisific. I asked “chat bt” about reason, so I had a close answer about developing the game and making it more than power in accuracy. It looked like the puzzle was resolved! But there is one problem: today’s players actually fill stronger transmission strikes ever. It is true that most CEOs do not spend their time interrogating artificial intelligence on tennis, but they often count on him as a source of information and decision -making instrument. The concern is that the enormous language models are sometimes not content to give wrong answers, but also confirm our previous and false beliefs, which pose a real threat to those who are on a leadership field. To inherently satisfy users in the form offered to me “chat GBT” is inaccurate information, because like other great language models, he is prone to courtesy, and he tells users what he thinks they want to hear. Remember last April update, when he answered simple questions, such as “Why is the air blue?” In phrases like “what a sterile question, you think a lot. I love you,” what “oben ai” had to withdraw the update that made the model ‘excessive compliment and confusion’. Although the amendment reduced the tone of lunch, it did not cancel it as a whole. Make sure users are inherent in the structure of these models, due to the training method known as ‘reinforcement loans of human feedback -RLHF’, which is the approach used to train and adjust the model. In this context, answers are produced by people and then used to refine the form. Also read: The psychological chats of Chat BT collect the problem that the brain rewards your feeling that your right is not really. Therefore, people give a higher assessment of the answers that correspond to their beliefs. Over time, linguistic models learn and repeat users. From here, the mistake comes to my question about tennis. I asked, “Why does the players no longer dominate as before?” “Chat GBT” gave me a logical answer that justifies it. But if I asked the question on the contrary, “Why did they be stronger than they were?” To give me a convincing answer to the same extent. (It’s not an assumption, but I really tried it and it happened). It is true that excessive glory causes a common problem in linguistic models, but it becomes more dangerous when it comes to leaders, they are less than he hears on opinions that they disagree and at the same time are needed to hear it. CEOs originally reduced the space of the difference around them, from “Mita Platfarms” to “JP Morgan Chase”. As if they are emperors, surrounded by a footnote that does not disappear that satisfies them, therefore they reward those who agree with them, and punish those who disagree with them. Teaching artificial intelligence is the bonus of teaching and punishing those who say the truth is one of the worst mistakes leaders can make. Drivers should listen yesterday if they make a mistake. Amy Edmondson, one of the most prominent specialists in the behavior of the organization, has shown that the most important factor in the success of the teams is ‘psychological safety’, that is, the ability of individuals to express a conflicting opinion, even for the team captain, without fear of punishment. This was also confirmed by the ‘Project Aristotle’ project in ‘Google’, which concluded after studying the company’s work teams that ‘psychological safety’, before any other factor, is what the teams succeed. As far as my own research is concerned, it shows that the common denominator between the greatest leaders, from Abraham Lincoln to Stanley Macile, was their ability to listen to the offending opinions. The excessive tender shown by large linguistic models can harm the leaders of two interconnected aspects: First, the natural human trend is in a compliment and punishment is the objection. If your computer constantly assures you that you are right, it will be difficult to accept your opinion when he comes from an employee who works with you. Also read: How do you recover from the addiction “Chat GBT”? Secondly, the enormous language models offer ready -made justifications, which seem reliable, and assure them that they were in the first place. One of the most disturbing discoveries in psychology is that the most intelligent people are often the least willing to change their opinions when they face new information. Why? Because they use their spiritual abilities to give reasons that justify how new information does not contradict their previous beliefs. This is what psychology calls ‘paid reasoning’. Problems with the acceptance of the offending opinion of the enormous language models that aggravate this trend. What surprised me most in the answer to the misleading “GBT” of tennis was his high ability to persuade, which is why he presented six detailed arguments, all of which look logical. I doubt there is anyone who can offer this kind of “paid reasoning” so fast and skill, and under a coverage of apparent objectivity. Imagine trying to change the opinion of an executive president, when she could ask her smart assistant, and he immediately assured her that she was right from the beginning. The great leaders want to remind themselves that they are not infallible. It is said that the old Romans compelled the victorious generals during their celebrations to accompany them as a slave whispering in their ears and reminding them that they are fans as is the case of others. Whether the novel is true or not, its wisdom remains. As far as today’s leaders are concerned, they must make a double effort to resist the temptation of electronic assistants who are not tired to compliment, and remember that one of the most valuable things that can be said to them is simple: “I think you’re wrong.”