Human & AI
Human & AI - Bridgeway Focus Technology

Human & AI

Everything we love about civilization is a product of intelligence, so the strengthening of our artificial intelligence can help civilization develop as never before - as long as we can keep the technology at its advantage. While sci - fi often presents AI as a robot with human characteristics, AI can cover everything from Google search algorithms to IBM Watson and autonomous weapons. Most researchers agree that a super - intelligent AI is unlikely to show human emotions such as love or hatred, and there is no reason to expect that AI will become deliberately mild or malicious. Because AI can be smarter than any human, we have no safe way to predict how it will behave. 

 However, it is becoming increasingly clear that AI algorithms embedded in omnipresent digital technology can encrypt social prejudices, spread conspiracy and communicate false news, strengthen the echo chambers of public opinion, divert our attention and even undermine our mental well - being. To effectively address these problems, a realistic concept of AI is needed, which is too often called "artificial minds ", on the exponential path to human thinking. Similarly, "intelligent" AI technology can inadvertently lead to "artificial stupidity" if it is poorly designed, implemented or adapted to the human social context. It is common to identify AI with machines that resemble human beings or simulate aspects of The human brain.

No alt text provided for this image

 

Data science products and AI applications are the most valuable when they are intelligently designed to meet the needs of end-users. Many AI systems can work on "autopilot" for a long time, but require human intervention in exceptional or ambiguous situations, requiring common sense or contextual understanding. 

 It is becoming increasingly clear that AI - based entertainment and social networking applications can affect human well - being in many ways.  Algorithms may also be used to increase human cognitive abilities - both System 1 "fast thinking" and System 2 "slow thinking" - as long as we adopt a man - centred approach to artificial intelligence - as long as we take a man - made approach to artificial intelligence. 

 If someone from a world of pure hunter - gatherers - - from a time when people were, more or less, simply another animal species - saw the vast human empires of 1750 with their towering churches, their ships crossing the ocean, their concept of being "inside ", and their vast mountain of being – what would they think of the intelligence of that community.

For example, the software and data behind Siri is AI, the voice of the woman we hear is the personification of such AI, and there is no robot at all. Ray Kurzweil then blurs things a little bit by defining singularity, as the Law on speeding up the return has reached such an extreme rate that technological progress is happening at an apparently infinite rate, and then we will live in a whole new world. Professor Linda Gottfredson describes intelligence as "a very general mental capacity, which includes the ability to reason, plan, solve problems, understand complex ideas, learn quickly and learn from experience. Artificial superintelligences vary from a computer that is a bit smarter than a human being to billions of times smarter computer. 

No alt text provided for this image


Research on AGI is an attempt to differentiate a thinking unit from today's AI technology, such as Machine Learning. People who "believe" in Singularity are happy to give the incredible power of post - Singularity because what happens next is quite unpredictable. Thanks to the many years of teamwork, we cannot prove that a 1, 000 line - up program cannot be compromised by external hackers, so we will certainly not be able to prove much about large AI systems at all. When people hear that automatic Learning is progressing at a high level and think about Learning machines in a new area, they tend to use them as a mental model in the way a person would learn this new domain. However, Machine Learning is very fragile and requires a lot of human preparation from researchers or engineers, special encoding for input data processing, special training data sets and a customized Learning structure for each new problem area. Today's computer - aided Learning is not at all a sponge as Learning human beings to engage in, making rapid progress in a new domain without having to be surgically modified or built with a purpose. 

 

No one in the technical world wants to face it, so we get lost in such AI fantasies. But if it could prevent the occurrence of AI in some way, it would have nothing to do with the real problem we fear, and it is sad, the hard thing we have to deal with. If AI stands for such mythology for the new creation we create, then it is just a stupid mess, which confuses everyone and destroys the future of the economy. Perhaps it is worse today because in the past we had at least the distinction between ethics and morality, where we could talk about two similar things, where one was a little more concerned with the mythologies of religion and one is a little less involved. 

要查看或添加评论,请登录

社区洞察

其他会员也浏览了