"AI Three Laws" is time to update

Robots must not harm human individuals, or see that human individuals will suffer danger and do not care about it.

The robot must obey the order given by people, when the command conflict with the first rhythm

Robots should protect their own survival as much as possible without violation of the first and second laws

In the middle of the 20th century, the three rules of robotics came out. In the pen of Asusmov, one of the three giants of the World Science and Fiction novels, the development of AI technology has been restricted to this frame and has affected it to this day.

No rules no standards.

In the science and technology world, the rules are transformed into the scientific and technological ethics represented by the three rules of robotics. If it is separated from ethics to the "freedom" of technology, the ending must cause anti -bump to human society. The movie "I, Robot" adapted from Asimov's novel "I, ROBOT" is such a story.

More than half a century has passed, and AI technology is already in the same day. What should the corresponding AI ethics be?

Recently, the 2022 World Artificial Intelligence Conference was held. As in previous years, AI ethics is one of the core issues.

During the meeting, IFF International Financial Forum, Singapore National University Lee Kuan Yew Public Public Public Public Public Public Public Public Public Public Public Public Public Policy Academy, Singapore AIII Artificial Institute of Artificial Intelligence International Research A number of international institutions and well -known scholars such as Intelligent Research Institute, Teslian, Shangtang Technology and other well -known scholars jointly launched the "Asian Science and Technology Promoting Sustainable Development Target Alliance", and the global human destiny community issued the "Shanghai Initiative", which includes: specifically:

1 overall goal: science and technology to promote the achievement of sustainable development of Asia

2 consensus: people -oriented, technical controlled

3 Class of Challenges: "Digital Economy Global", "" Supply Chain Gully "and" ESG Honggou "

Four measures: Digital infrastructure interconnection, digital public service tolerance development, digital trust and ethical security, digital social capacity building

Facing the risks that emerging technology may bring, the three laws of AI are time to update.

Refresh AI ethics, it is imperative

Is the first three rules of robotics applicable? The artificial intelligence industry often argues for this.

The answer is yes, because the three rules are the foundation of AI ethics. But on this basis, it should be extended with the times.

According to data from the Ministry of Industry and Information Technology, as of now, the scale of China's artificial intelligence core industry has exceeded 400 billion yuan, and the number of enterprises exceeds 3,000. And with the accelerated integration of artificial intelligence with manufacturing, transportation, medical care, people's livelihood and other fields, the industrial scale will continue to erupt.

In the context of the outbreak of the industry, the construction process of the global AI ethics system has accelerated significantly in recent years.

For example, in 2016, Google, Facebook, IBM, Amazon, Microsoft and other well -known American technology companies jointly announced the establishment of the "Partnership on AI" cooperation organization and released the "Principles of PARTNERSHIP on AI". Conducts that need to be complied with.

Back to China, the exploration of enterprises and the guidance of policies are being launched simultaneously.

Taking Shangtang Technology as an example, at the World Artificial Intelligence Conference last year, the AI ​​ethics concept of "balanced development" was officially proposed, and the three core principles of "people -oriented, technical controllable, and sustainable development". Based on this, we have carried out practice from many aspects such as organizational construction, mechanism construction, tool construction, cultural construction, and ecological construction.

Summarizing experience in practice, at the conference this year, Shangtang Technology once again emphasized in the "Balanced Development" artificial intelligence governance white paper "that" the development and governance of artificial intelligence should be parallel. 'And' steering wheel ', complement each other, lack one ", and further propose the development of" responsible and evaluated "artificial intelligence.

In terms of policy, the ethics and governance of technology also gives methodology.

In March of this year, the General Office of the Central Committee of the Communist Party of China and the General Office of the State Council issued the "Opinions on Strengthening the Governance of Scientific and Technological Ethics", which clearly mentioned two points, ethics first, and based on national conditions. The former emphasizes the importance of scientific and technological ethics in the guarantee of scientific research innovation; the latter requires the construction of the ethical system and the process of governance to be open and tolerant, adapt to local conditions, and benefit the people.

Aspiring to contribute to the global scientific and technological ethics governance of China's practical wisdom and Asian consensus plan, the significance of the establishment of Asian science and technology to promote the sustainable development target alliance is prominent.

As AI industrialization will usher in a large -scale outbreak, there must be ethical principles and the direction of the ESG principle. It is imperative to use the alliance to gather industrial forces and jointly refresh the current AI ethics law.

The bottom line and pursuit, let AI balanced development

The complete ethical governance system is the standard for promoting the sustainable development of scientific and technological innovation. After identifying whether it is a question, let's analyze the specific principles.

The first thing that can be determined is that the bottom line of all technologies is people -oriented, and artificial intelligence is not enough.

In the last century, Asimov's foundation for AI ethics, the core idea was people -oriented, so the robot was required not to harm humans and obey human instructions.

From novels into reality, the rooting of AI has made the scope of the application of ethics more complicated.

Big data, computer vision, voice recognition, machine learning, etc. The popularization of a certain technology must face up to the development differences in the history, culture, society, and economy of different countries and regions, so that technology can be used to achieve no prejudice applications. Previously, there were previously available. The company's image recognition algorithm is the most typical negative case that is recognized by black people as "spiritual long animals". Of course, many positive samples can be found in the industry. For example, parents in modern society will have a headache because of too much contact with the electronic screen. The "meta radish" AI chess robot recently released by Shangtang Technology has found a good solution for the child's interest transfer. Relying on AI visual technology and robotic arm technology, let the chess robots interact with children, which can effectively exercise children's thinking ability.

Secondly, in the ultimate direction, the development of AI technology cannot deviate from the orbit that promotes the sustainable development of society, economy, culture, and environment.

This is also an industrial consensus, which can be seen from the name of "Asian Science and Technology Promoting Sustainable Development Target Alliance". In the "Opinions on Strengthening the Governance of Science and Technology Ethics", the first ethics principles clearly stated that scientific and technological activities should "enhance human well -being" and promote the peaceful development of human society and sustainable development.

Looking at it further, sustainable development covers various aspects such as society, economy, environment, and culture. The technology needs to resonate with the times as the times, and to help the sustainable development according to the facts of various industries.

In terms of environmental protection, in recent years, the world has been advancing to the green target of carbon neutralization, and AI technology has greatly promoted. For example, as the world's major source of carbon dioxide emissions, the digital transformation and cost reduction of the industry can effectively promote emission reduction. To this end, Shangtang Technology combines AI and AR technology to tailor the power grid intelligent inspection solution for the power industry, and use the assistance of intelligent technology to save operation and maintenance time and labor cost. Sustainable development has added a force.

Finally, the question that needs to be considered is how to ensure the above -mentioned bottom line and pursuit from planning to implementation as scheduled in the process of technology application? The answer is simple -technical controlled.

Technology is controllable. In fact, it is necessary to ensure that artificial intelligence is developed by human beings, serving human beings, and controlled by humans. Correspondingly, its ethical responsibility caused by artificial intelligence applications should also be borne by its controller (human).

The clarification of the responsible subject has a security lock of ethics and rule of law to the application of technology, which is more conducive to controlling the risk of technology development to the source. Following the products developed by technical controlled concepts, a series of problems caused by out of control are naturally avoided.

In summary, combined with the actual cases of the application of AI technology, from the perspective of the underlying logic of industrial development, the people -oriented, technical controlled, and sustainable development, Shangtang Technology's ethical principles based on "balanced development" are precisely adapting to the current cycle. "New AI Law".

The current governance is to "treat the disease" in the future

Google's motto is believed to be familiar and not evil. As a technology company, including Google, the industry knows the harm of technology. The existence of scientific and technological ethics is to eliminate the possibility of potential evil.

The industry often gives an example: Imagine that one day in the future, you go to the airport in a hurry, so you send an instruction to the autonomous driving car "Send me to the airport, the faster the better", and AI is speeding at speed when performing the instruction speed. , Through the red light and knocked down the pedestrian.

This is the disastrous consequences of the lack of scientific and technological ethics. In order to prevent this scene, enterprises must give governance in advance during the technological research and development and product design stage.

Go to medical treatment, heal the disease.

In this regard, Yang Fan, a co -founder of Shangtang Technology, Yang Fan, member of the Artificial Intelligence Ethics and Governance Committee, said: "The boundary and core requirements of the AI ​​ethics risk are the responsible and evaluated artificial intelligence. One step, do one step first. "

The core feature of "Responsible and evaluated artificial intelligence"

Source: Shangtang's "Artificial Intelligence Governance White Paper"

Responsible and evaluated. After clarifying the subject of responsibility, Shangtang Technology put forward further requirements for artificial intelligence governance. Specifically include seven aspects: responsibility for people, responsibility for society, environmental responsibility, traceability of responsibility, risk assessment, process assessment, and effectiveness.

Through the improvement of the governance system and the past governance practice of Shang Tang, we can see that artificial intelligence, the cutting -edge technology for the star of the Star, must be down -to -earth every step.

This is the point of view of French philosopher Sartre, "freedom is a responsibility."

Under the outbreak of the industry, the freedom of the development of AI technology is actually based on the responsibility of being responsible for themselves, to others, society, and technology itself. Only in this way can the value of technology be maximized. Over time, the benefits of scientific and technological ethics construction and governance work will become more prominent.

But having said that, scientific and technological ethical governance is by no means easy.

The three laws of robotics set by Asimov are often broken in science fiction works. The more complicated real life, the corresponding challenges can be imagined. Moreover, this is the common propositions of the global and industry, and participants in all industry are needed.

Fortunately, someone is insisting on doing difficult and correct things.

Author of this article: Industry Fenghuihui

The industry is unpredictable.

- END -