The 40th International Conference of the Data Protection and Privacy Specialist

(40th,

2018.10.22-26, Brussels)

Chinese translation

Wu Shenkuo, Zhou Jie, Yang Yingying/Translation

Beijing Normal University International Research Center for Internet Criminal Law

October 24, 2018

Organizer:

•The French National Commission for Information and Freedom (CNIL)

•European Data Protection Specialist

• Italian Personal Data Protection Specialist

Supporter:

•Argentina News Agency

•Quebec Information Commission, Canada

• Norwegian Data Supervision Commission

• UK Information Commissioner's Office (ICO)

• Swiss Federal Office of Data Protection and Transparency

•Belgian Data Protection Agency

• Hong Kong Personal Data Privacy Specialist

•Ireland Data Protection Commission

•Polish Office of Data Protection

• Mexico National Institute of Transparency, Information Access and Personal Data Protection (INAI)

• Hungary National Data Protection and Freedom of Information Administration

•Germany Federal Specialist in Data Protection and Information Freedom

•Office of the Canadian Privacy Commissioner (OPC)

Considering the preliminary discussion on artificial intelligence, robotics, privacy and data protection at the 38th International Conference of Data Protection and Privacy held in Marrakech;

Recognize that AI systems can bring significant benefits to users and society, including: increasing the speed and support of relevant decision-making in relevant processes; creating new ways to participate in democratic processes; improving the efficiency of public sectors and industrial enterprises; and in public health , health care, safety, sustainable development, agriculture and logistics, provide new methods and solutions in various fields such as health care, safety, sustainable development, agriculture and logistics; bring new opportunities to scientific research and education; and provide more personalized services to individuals.

Given the significant progress of artificial intelligence in certain areas, especially in the processing of large amounts of information, the analysis and prediction of human behavior and characteristics, and related fields such as robotics, computer vision and automated systems, significant progress may be made in the near future;

The rapid development of big data and artificial intelligence, especially machine learning, and in particular deep learning technologies, allows algorithms to solve complex operations that lead to potential decisions, but this makes these processes more opaque;

Refuting respect for privacy and data protection is increasingly challenged by the development of artificial intelligence, which requires complementary to ethical and human rights considerations;

Considering that machine learning technologies, especially artificial intelligence systems, may rely on processing large amounts of personal data for development, which may affect the protection of data and privacy; also considering the potential risks brought about by the current market concentration trend in the field of artificial intelligence;

Recognize the link between the collection, use and disclosure of personal information – on the one hand, traditional privacy and data protection issues, and on the other hand, direct impact on the wider human rights, most notably discrimination, freedom of speech and Freedom of information. Therefore, the recognition of data protection and privacy rights requires that authorities consider human rights more broadly and should cooperate with other agencies dealing with human rights;

It is pointed out that some data sets based on machine learning and artificial intelligence systems used for training have been found to contain inherent biases that lead to decision-making that may unfairly discriminate against certain individuals or groups, which may limit the availability of certain services or content, thereby interfering with individuals Rights, such as affecting freedom of speech or causing people to be excluded from certain aspects of their personal, social, and professional life;

Emphasizing that AI-driven systems that cannot explain decisions raise not only basic issues of privacy and data protection laws, but also issues of responsibility in the event of errors and harm;

Noting that many stakeholders in the field of AI have expressed their concerns about the risks of malicious use of AI and risks associated with privacy, data protection and human dignity, and point to the development of AI. Combined with large-scale surveillance, raises concerns that it may be used to restrict fundamental rights and freedoms;

Emphasizing that these risks and challenges may affect individuals and society, the extent and nature of potential consequences of harm are currently uncertain;

Emphasize the importance of trust because strong data protection and privacy protection help build individual trust in how data is processed, thereby encouraging data sharing and thus promoting innovation;

It is believed that the current challenges posed by the development of artificial intelligence and machine learning systems reinforce the need to adopt international methods and standards to ensure that human rights are promoted and protected in digital development at the international level;

Reaffirm the commitment of the Data Protection Agency and the International Conference of the Data Protection and Privacy Commissioner to maintain data protection and privacy principles to adapt to the changing environment at present, especially by attracting resources and developing new skills, to prepare for future changes.

The 40th International Conference of the Data Protection and Privacy Commissioner believes that the creation, development and use of any artificial intelligence system should fully respect human rights, especially the right to protect personal data and privacy and the right to human dignity from damage, and should provide a solution. Programs that enable individuals to control and understand artificial intelligence systems.

Therefore, the meeting agreed to the following principles as the core values ​​for protecting human rights in the development of artificial intelligence:

1. Artificial intelligence and machine learning technologies should be designed, developed and used at the basic human rights level in accordance with the principle of fairness, especially:

Consider individual reasonable expectations by ensuring that the use of artificial intelligence systems is consistent with their original purpose and that the data is used in a manner consistent with the original purpose of its collection;

It is necessary to consider not only the impact of the use of artificial intelligence on individuals, but also its collective impact on groups and society as a whole;

Ensure that artificial intelligence systems are developed in a way that promotes human development without hindering or harming humans, thus recognizing the need to divide and define boundaries for certain uses;

2. Continued attention and vigilance on the potential impact and consequences of artificial intelligence systems and the establishment of accountability, especially:

Promote the ongoing monitoring and impact assessment of AI systems by individuals, regulatory authorities and other third parties as stakeholders and the responsibility for periodic review of oversight mechanisms, including through audits;

Promote collective and shared responsibilities covering the entire chain of participants and stakeholders, including coordinating the best ways to develop and share standards;

Increase investment in public awareness, education, research and training to ensure that understanding of information and artificial intelligence and their potential impact in society remain at a good level;

Establish proven governance processes for all relevant actors, such as by relying on trusted third parties or establishing independent ethics committees.

3. The transparency and intelligibility of artificial intelligence systems should be improved to achieve effective implementation, especially:

Increase investment in public and private scientific research on artificial intelligence;

Improve transparency, intelligibility and accessibility, such as by developing innovative communication methods, taking into account the different transparency and information needed by each relevant audience;

Make organizational practices more transparent, especially by increasing algorithm transparency and system auditability while ensuring the meaning of the information provided, and

To guarantee the right of self-determination of information, especially by ensuring that individuals always receive appropriate information when interacting directly with artificial intelligence systems or when they provide personal data processed by such systems,

Provide adequate information about the purpose and effectiveness of artificial intelligence systems to verify continuous consistency with individual expectations and enable these systems to achieve comprehensive human control.

4. As part of the overall “design ethics” approach, AI systems should be designed and developed responsibly by applying privacy principles and design privacy by default, especially:

Implementing technical and organizational measures and procedures—as proportional to the type of system developed—ensures that the privacy and personal data of the data subject are respected when determining the processing methods and data processing,

Evaluate and document the expected impact of an AI project on individuals and society at the beginning and related developments throughout the life cycle,

Determine specific requirements for the moral, reasonable use of the system and respect for human rights as part of the development and operation of any artificial intelligence system,

5. Empowering everyone should be promoted and the exercise of individual rights should be encouraged to create opportunities for public participation, especially:

Respect data protection and privacy rights, including applicable information rights, access rights, opposition to processing and deletion rights, and promoting the exercise of these rights through educational and advocacy activities,

Respect relevant rights, including freedom of speech and information, to expand the right to object or to inform the various technologies that affect individual development or opinions, and where significant impacts exist, the individual shall be guaranteed, where applicable, not subject to automatic processing solely the right to decide, while ensuring that individuals have the right to challenge such decisions,

Utilize the capabilities of AI systems to promote equal empowerment and enhance public participation, such as leveraging adaptive interfaces and accessible tools.

6. Illegal bias or discrimination that may arise from the use of artificial intelligence data should be reduced and mitigated, including:

Ensure respect for international legal instruments on human rights and without discrimination,

Invest in technical methods to identify, address and mitigate bias,

Take reasonable steps to ensure that the personal data and information used in automated decision-making is accurate, up-to-date and as complete as possible,

Develop specific guidance programs and principles to address bias and discrimination and raise awareness among individuals and stakeholders.

Taking into account the above principles, the 40th International Conference of the Data Protection and Privacy Commissioner called for the development of common governance principles on artificial intelligence and promote coordinated efforts in this area by countries to ensure that the development and use of artificial intelligence is in compliance with norms, ethics and humanity. Values, respect human dignity. These shared governance principles must be able to solve the challenges brought about by the rapid development of artificial intelligence technology and solve all cross-sectional problems based on the multi-stakeholder approach. It must be promoted globally because the development of artificial intelligence is a transnational phenomenon that may affect all humanity. The General Assembly should also be an integral part of the international community's efforts to cooperate with and provide support to industry or general authorities in other areas such as competition, market and consumer regulation.

Therefore, as a contribution to future common governance at the international level, and in order to further formulate guidance plans to supplement the principles of artificial intelligence ethics and data protection, the 40th International Conference of the Data Protection and Privacy Commissioner has established a response to the challenges brought about by the development of artificial intelligence. permanent working group.

The AI ​​Ethics and Data Protection Working Group will be responsible for promoting the understanding and respect of the principles of this resolution by all parties involved in the development of AI systems, including governments and public institutions, AI system designers, providers and researchers, enterprises, citizens and end users of artificial intelligence systems.

The Working Group on the Ethics and Data Protection of Artificial Intelligence should take into account the work carried out by other working groups of the conference and should report their activities to the General Assembly regularly. Therefore, this Congress fully supports the active public debate on digital ethics, helping to build a strong moral culture and personal awareness in this area.