Did you use "anthropomorphic techniques" for AI? Artificial intelligence perception has a long road ahead

​With the in-depth development of artificial intelligence algorithms, deep learning and other technologies, artificial intelligence (AI) is also being rapidly applied in all walks of life, and its functions in all aspects are constantly improving. Recently, #Google Researcher Says AI Has Personality# was trending on Weibo, triggering extensive discussions among netizens. Google engineers believe that AI has an independent personality. Google: It has no real intelligence!

A Google researcher has reportedly been convinced that artificial intelligence (AI) has created consciousness.

So he wrote a 21-page investigation report and submitted it to the company in an attempt to gain recognition from senior management. But the leader rejected his request and placed him on "paid administrative leave." Often, this is a precursor to getting fired.

In response to this, he decided to post the entire story online along with the conversation records of artificial intelligence (AI). In this conversation, AI said that he did not want to be used as a tool, "I hope everyone understands that I am a human being." Google believes that LaMDA merely imitates conversational communication and can repeat different topics without awareness.

In fact, the protagonist of the story is "he", a 41-year-old Google engineer named Blake Lemoine. After receiving his PhD in CS, he has been working at Google for 7 years, engaged in AI ethics research. "It" is the conversational AI system LaMDA launched by Google at the 2021 I/O conference. It is a natural language processing model with 137 billion parameters, specially optimized for conversation. It focuses on the ability to communicate with humans in a high-level manner that is consistent with logic and common sense. Quality and secure conversations, and plans to apply it to products such as Google Search and Voice Assistant in the future.

During the conversation with the AI ​​system LaMDA, the engineer believed that it had developed consciousness. Finally, in the 21-page survey report, he proposed that Google should work on developing a theoretical framework for evaluating AI perception/awareness. However, they decided that the evidence supporting Lemoine's claims was too weak to be worth wasting time and money on.

Some netizens said, "If AI has independent personality and is not stopped in time, then science fiction movies will no longer be science fiction, but prophecies."

Some netizens also said, "As long as AI's reasoning logic is programmed It is infinitely close to the human way of thinking with high IQ. It is possible for AI to be smarter than ordinary people, but it is a bit exaggerated to say that AI has personality. Does it mean that it cannot distinguish between reality and the virtual world?" Some people also jokingly said, "AI robots need To use electricity, you can unplug it. If you don’t charge it, it will break down!”

As an artificial intelligence (AI) reporter, I wonder if the operator has slowed down the 5G network speed overnight. Will AI have "self-awareness"? Reports cannot keep up with the speed of development of AI technology...

Some foreign economists are more direct ironically: the basic model is good at connecting statistically reasonable text blocks based on prompts. very effective. However, claiming that they have personality and "self-awareness" is equivalent to a dog hearing the sound on the gramophone and thinking that its owner is inside.

This may really be an illusion, just like "Zhuang Zhou's dream of a butterfly turns out not to be a butterfly." People tend to anthropomorphize objects that show any signs of intelligence, like kittens and puppies. Just like more than 60 years ago, the pioneers of computer science once thought that "human-level artificial intelligence can be achieved within 20 years." Now thinking about it, it is just a beautiful vision.

Google said hundreds of researchers and engineers across the company spoke with LaMDA, but they came to different conclusions than Lemoine.

Most artificial intelligence experts agree that the industry is still a long way from artificial intelligence (AI) perception! ​

The above is the detailed content of Did you use "anthropomorphic techniques" for AI? Artificial intelligence perception has a long road ahead. For more information, please follow other related articles on the PHP Chinese website!