Based on the current authoritative reports, the author believes that access in the field of government services mainly presents two modes: internal service and external service. Internal services focus on using auxiliary government staff and government system technology research and development and management teams to improve their work efficiency and quality; while external services are directly applied to serve the general public, and are vividly called "AI civil servants". Table 1 lists the service objects, service content, service goals and specific service cases of these two models in detail. Through classification and sorting, the wide applicability in different application scenarios is clearly demonstrated.

Table 1 Classification of service modes for accessing government affairs system

Note: The content of the form is compiled and prepared by the author.

Access to government services

Ethical Challenges

It can be seen that whether it is in-depth empowerment within the government or convenient services that directly benefit the people, it has undoubtedly opened up unprecedented broad opportunities for the digital transformation of government departments. This trend has been vividly reflected in Binyang County, Guangxi. The county party secretary solemnly emphasized at an important official meeting that all participants must actively learn and master the usage methods, which fully reflects the local government's firm determination and positive actions to embrace intelligent technology and accelerate digital transformation. However, as local governments scramble to join the embrace trend, the ethical challenges of accessing government affairs systems are also worthy of attention. The author believes that there are the following ethical challenges to accessing the government system:

Issues of Definition of Responsibility under Human-machine Cooperation

Although access to the government affairs system has reduced the burden on government staff, it has also caused concerns about the issue of defining responsibility under human-computer cooperation. For example, decision recommendations are provided based on massive data and algorithmic models, which may be biased, resulting in unreasonable or errors in decision making. How should responsibilities be defined in this case? For example, when the government affairs system is connected to the government affairs system, it makes mistakes when approving or handling a specific matter, how should responsibilities be divided? In response to this, Gao Zeng, deputy director of the Government Services and Data Management Bureau of Futian District, Shenzhen, emphasized in an interview that "the accessed "AI digital intelligence employees" cannot make decisions alone, but work is carried out under the leadership of the guardian, and the guardian is responsible." However, it is difficult to ensure that all links can be artificially supervised in specific practice, and once errors occur in some special government scenarios, the losses caused are huge, and it is not very reasonable to attribute them all to the guardian. Therefore, how to clearly define the responsibilities of various behaviors in the deployment of government service systems under human-computer collaboration is an ethical issue that needs to be given priority in the current and in the future.

Algorithm bias and discrimination issues

Algorithm bias and discrimination are the primary issues that are considered in the current practical application of artificial intelligence, and this issue should be taken more seriously in the field of government services. As the provider of public services and the subject of social management, the government's decisions and actions directly affect the interests and well-being of the general public. The "Aro Impossible Theorem" in economics clearly mentions that in a society, if everyone has different preferences, it is impossible to achieve a common choice that everyone can accept through voting or other social selection mechanisms. That is to say, citizens’ needs and preferences are not necessarily the same in practice, and AI algorithms may be difficult to balance the interests of everyone. For example, in resource allocation or policy making, AI may tend to meet the needs of the majority and ignore the rights and interests of minorities.

Efficiency and value balance

In recent years, improving government administrative efficiency has become a priority for government reforms in many countries. For example, at the suggestion of entrepreneur Musk, the Trump administration in the United States established the "Doctor of Government Efficiency" (DOGE) at the beginning of this year, and used AI technology to conduct a series of explorations in improving administrative efficiency, controlling and reducing government improper spending. Similarly, governments across China actively connect AI technologies to government service systems, aiming to use AI technology to improve administrative efficiency and reduce the work burden of government personnel, so as to better serve the general public. However, excessive reliance on AI technology may bring about the problem of "efficiency first, value lag". For example, AI will automatically and quickly process (pass or reject) a large number of service window application services in accordance with algorithmic instructions. Although this efficient processing significantly improves administrative efficiency, it may also ignore individual differences and special circumstances, resulting in a lack of humanistic care at the value level of public services.

Data security and privacy protection

Unlike other single subjects' data security and privacy protection, government data involves massive data from multiple subjects such as citizens and enterprises, and has the characteristics of high sensitivity and public interest attributes. Their security is directly related to national security and citizen rights. Once the government system is attacked by a cyber attack or system vulnerabilities are maliciously exploited, the attacker may manipulate the government system's Q&A mode through simple instructions, which will lead to a large amount of citizen privacy data leaks. This risk not only undermines public trust in the government, but can also cause serious social problems and even threaten national security.

Government trust issues

The above potential ethical concerns will lead to a more serious problem, namely the issue of citizens' trust in the government. For example, algorithmic bias and discrimination can lead to unfair treatment of certain groups in government services, which directly undermines citizens' perception of government fairness. For example, if certain privacy data of citizens or enterprises are improperly acquired or accidentally leaked in government service systems supported by AI technology, it will also damage the credibility of the government. These potential ethical issues may weaken and damage citizens' "trust inertia" in the government, and thus pose severe challenges to the government's future public governance.

How to deal with:

"Dual rationality wisdom governance"

"It is better to stay happy for a long time than to be happy when you see it." In response to the above potential ethical concerns, the author believes that government departments should not only seize this rare AI opportunity, but also maintain sufficient rationality and strategic determination in the process of embracing it, and make good use of it from a long-term perspective. On the one hand, government departments should promote the transformation of instrumental rationality of "efficiency first" to the dual rational concept of "balance of efficiency and value", and strive to pursue efficiency while not neglecting the adherence to ethical values. On the other hand, given the rapid development of artificial intelligence technology, the construction and improvement of the ethical and legal system need to be followed closely, and with the development concept of "intelligent governance and peer-to-peer" to ensure that every step of technology can move forward steadily within the legal framework. As Prime Minister Li Qiang clearly stated in the dialogue between Davos Forum (World Economic Forum) at the Davos Forum (World Economic Forum) early last year, "We must not only use AI, but also manage AI." At the National People's Congress and the Chinese People's Political Consultative Conference held on March 8, the world's first systematic AI ethics law, "Artificial Intelligence Ethical Governance Framework", was passed, which further strengthened the orientation of science and technology to be good and intelligent governance.

While the law follows up, the development of technology has not stopped. On March 6 just passed, the Chinese team officially released the general AI agent "Manus", which once again triggered the capital market and public opinion market, and related concept stocks hit the daily limit one after another. In other words, China's AI technology has not stopped, and this is exactly the beginning. With the continuous evolution of AI technology, more AI products will emerge in the market.

All the past is the prologue. Everything in the future has expectations. In short, the author believes that in the future, government departments should keep up with the pace of the times and explore in depth how to take into account the protection of ethical values ​​in the process of embracing technological innovation. Only by doing a good job in this proposition that needs to be paid attention to in the current and future period can we ensure that science and technology always be good and move forward steadily!