The "AI Now" seminar held in July this year is the last link in a series of studies jointly promoted by the White House Science and Technology Policy Office and the U.S. National Economic Council. The previous series of studies have analyzed AI from different perspectives, from policies and regulations to AI security control, to AI charity and how to tap AI's more potential. This "Aowow" mainly discusses the influence of AI in social and economic fields in the next decade. Experts and scholars from various fields in the world gathered to express their opinions. The issues discussed include: What are the problems caused by the rapid development of AI at this stage? How to better understand and use AI to create a fairer and fairer future?
There are numerous social and economic problems. This “AI Now†mainly discusses the principles of “medical careâ€, “labor employmentâ€, “AI fairness†and “AI ethicsâ€.
The reason why “medical†and “employee employment†were chosen as the main topic is that currently the AI ​​infiltrate into the two fields widely, and the problems that AI can bring are more prominent and obvious in these two fields. "AI fairness" and "AI morality" are the questions that everyone is concerned about in the future: Will AI help the world's majorities or will increase social inequality? And how to ensure that the benefits of AI are enjoyed by all humans?
The seminar was held to make AI better for the benefit of human society. Through numerous ways for experts and scholars to gather for discussion, this “AI Now†seminar has significant implications both inside and outside the artificial intelligence academic world.
Questions and Suggestions
The seminar foresees possible future AI situations and gives corresponding recommendations. What needs to be stated is that the following suggestions incorporate the wisdom of all participants and do not represent the position of an individual or an organization.
As AI is increasingly applied to all aspects of socio-economic life, the following list of issues and corresponding recommendations can be used as reference guides for investors and practitioners in related fields.
1. Problems: The development and application of AI depend on specific infrastructure and human and material resources. The shortage of these basic resources will undoubtedly limit the development of AI, and the mastery of these infrastructures and resources will be crucial in the early development of AI.
Recommendations: Improve the resource base for the development of AI from multiple sources. Focus on the construction of supporting areas such as data sets, computers, and related personnel education and training.
2. Question: Although the current level of AI is still at the primary level, AI has existed as an artificially assisted role in many areas and has had an impact on labor relations. Jason Furman, president of the Obama Economic Advisory Council, said that low-tech manual labor is the most likely position to be replaced by AI and automated machinery. If robots begin to compete with humans, the distribution of human resources will also usher in change.
Suggestions: Update your thinking and skills to deal with the changes in the employment structure brought about by AI participation. Future AI machines will assume most of the low-skilled jobs, and people will need to adjust their skills reserves and revenue and expenditure to cope with the new situation.
3. Problems: The process of AI and automation is usually performed behind the scenes that people can't see. Without human participation, the machine may make decisions that are unfair or careless. As AI applications grow further, AI judgments and errata will become more important and more difficult.
Recommendation: Support AI calibration and errata research, and AI error hazard assessment procedures should also be put on the agenda. These studies should be matched with the rapid development of AI, just as the administration of justice in the human system. This will promptly discover the mistakes made by AI and avoid serious consequences.
4. Issues: Research on fairness and accountability of public and private institutions under the AI ​​model seems to be in conflict with current US laws, such as the Computer Fraud and Abuse Act (CFAA) and the Digital Millennium Copyright Act (DMCA).
Suggestions: What needs clarification is that neither the Computer Fraud and Abuse Act nor the Digital Millennium Copyright Act restricts research.
5. Problems: Although AI is being used at a rapid rate in many fields such as medical care and labor, there is currently no universally accepted method for assessing the impact of AI.
Recommendation: Support the study of the AI ​​impact assessment system. And research in this area should work with government agencies to make the results available to government administration.
6. Questions: Those whose rights have been harmed by the deployment of AI are often ignored.
Recommendation: When building an AI system, the opinions of the affected people should be heard. AI should be designed by all parties to avoid being unfair and too aggressive.
7. Problems: The research of AI is mainly focused on electronic technology. Attention to human nature issues is often insufficient. In the future, the members of the field of computer science will be characterized by homogeneity and simplification, which is detrimental to the vision and experience of AI developers, and then affects the creation of AI products.
Recommendations: AI researchers and developers should be as diverse as possible, and the diversity and diversity of developers will also bring more abundant AI products. In the future, the AI ​​field should support interdisciplinary research in many ways, so that the AI ​​system can integrate electronic computing, social science, and humanistic atmosphere.
8. Question: Existing ethical guidelines can no longer cope with the complexity of AI's problems in reality. (For example, in medical treatment, law enforcement, criminal judgments, and labor services, etc.) At the same time, in the computer class in universities, although these polytechnic courses have gradually begun to pay attention to moral education, they have not been fully implemented in practice.
Suggestions: Collaborate with professional organizations such as AAAI, ACM, and IEEE to promote ethical standards that can meet new situations. At the same time, implement the education of these new moral principles in the school classroom. Every student who is interested in computer science should also receive moral education such as civil rights and freedom in addition to professional classes. Correspondingly, practitioners in areas where AI is infiltrated (such as medical facilities) should also be aware of these new ethical standards.