AI this, AI that, everything is all about automation now. Every tech giant is pouring billions into the development of advanced AI that can solve unorthodox problems. Moreover, they don’t pass up on any chance to preach how AI will change the future. But a new report published by WSJ begs to differ from the idea.
It gives an insight into how AI systems of today are merely a gimmick and nothing more. Researchers believe that AI companies are overexaggerating the benefits and possibilities of current-gen AI tech. They believe that it is against the ethics to promote or highlight such possibilities from AI which may never come to fruition.
Why researchers are angry about the claims about AI?
Last week, a Google engineer made claims about the LAMDA AI had become sentient. The claim was then proved wrong by Google’s research team. It ended in an abrupt suspension for the employee. Researchers believe that AI hasn’t reached the level of consciousness yet. Even if becomes possible, it will take decades to reach that goal.
As per the WSJ report, researchers claim that the AI tech companies have been building, is to perfect their own products. Tasks like mapping users’ preferences into buyer personas, relevant ad targeting, and ad recommendations are all that the AI systems by big tech are capable of.
Moreover, these big tech companies have amassed huge profits by tweaking their AI systems. But the claims that they can achieve consciousness and think like humans is still untrue.
Thinking like humans is certainly not a good trait for an AI. Google’s research employees discovered that thinking like humans would force AI to have the same potential for failure. So, they would become more prone to erroneous decisions, some of which could endanger lives.
They mentioned a case of misinterpretation by Facebook AI which led to a wrongful arrest of a person. He posted a good morning greeting in Hebrew which the AI misinterpreted as ‘hurt them’ in English. The end result was a swift arrest made by the local police which was obviously wrong. These kinds of examples show that AI is far from what the companies are trying to sell in public forums.