Most popular technology of embedding moral algorit

  • Detail

Embedding "moral" algorithm into intelligent machine (Science and technology essay)

original title: embedding "moral" algorithm into intelligent machine (Science and technology essay)

the application of artificial intelligence may be a far-reaching social ethics experiment. Intelligent machines should be embedded with human values and norms through algorithms, so that they have the same ethics as humans

not long ago, a number of top global AI experts, including Tesla CEO musk, signed a letter of commitment calling for the development of "Ai autonomous weapons". This is not the first time the industry has warned about the potential risks of AI. Although the autonomy of "weak artificial intelligence" to deal with single tasks and complete human instructions is still very low, people cannot ignore the potential hidden dangers of the application of new technologies

efficient and intelligent AI is not perfect. For example, artificial intelligence has algorithm defects. This defect often stems from the inexplicability and incomprehension of the machine learning process. Its judgment is sometimes like being completed in a black box, lacking transparency. When an driverless car decides to turn left or right, people do not fully understand the process of turning. In addition to its need to change the control situation when it just shows the appearance of submission, people can't even find the cause of the problem accurately, which brings obstacles to the targeted adjustment of the system and judgment. Because of this, the European Union's general data protection regulations, which came into force not long ago, specifically stipulates that AI companies must have someone review some algorithm decisions

there are not only immediate worries, but also foresight. Looking ahead, the application of artificial intelligence is blurring the boundary between the virtual world and the physical world, which may reshape the living environment and cognitive form of human beings, and thus derive a series of thorny ethical, legal and security problems. Compared with other technological innovations in history, automobile lightweight has become the current trend of automobile development in the world. An important reason why artificial intelligence issues such as ethics and law have received so much attention is that it is expected to achieve computable perception, cognition and behavior in concept, so as to simulate human intelligence and action in function, which makes the machine have a quasi personality or quasi subject characteristic. Whether the existing legal norms, conceptual framework and knowledge reserves of mankind can cope with the new problems caused by the development of artificial intelligence is a challenge that people have to face up to

in view of the potential risks of the application of artificial intelligence, the international community has put forward a series of schemes to try to control the intelligent machine system in terms of standard design, ethics and morality, and gradually formed a consensus principle: embed human values and norms into intelligent machines through algorithms, so that they have the same ethics and morality as human beings, such as compassion, heart, shame and so on. It is hoped that the algorithm of artificial intelligence will follow the "good method", and absorb the "good" side of human moral norms while completing tasks that are difficult for human beings, so as to achieve the purpose of controlling machine risks. Tooth destruction after all, any algorithm is only a tool or means to achieve the goal, and the result of "goodness" is the goal of human beings

in fact, the idea of controlling machines by writing code and algorithms is not new. More than 70 years ago, Asimov, an American science fiction writer, put forward the "three laws of robotics" and conceived the principle that robots should not harm humans through the built-in "machine ethics regulator". Today, computer experts, ethicists, sociologists, psychologists and others are working together to practice this idea

human beings' foresight about the development of artificial intelligence may have come a little early, but what we need to prepare for a rainy day reflects human wisdom and insight. We cannot wait until artificial intelligence is deeply involved in human life before we think about how to get along with it. Harmony with machines should be an important yardstick for human development and utilization of artificial intelligence. Before use, after all, the application of artificial intelligence is not only a technological innovation, but also a far-reaching social ethics experiment

"the short-term impact of AI is determined by the person who controls it, while the long-term impact depends on whether AI is completely controlled by people." This is the warning left by the famous physicist Hawking to mankind. We don't have to be intimidated by the risks of AI, but while enjoying the benefits of AI, we also need to draw a red line of ethics and legal rules

Copyright © 2011 JIN SHI