769 views
Sber developed AI ethics principles
Head of Sber German Gref noted that the principles were elaborated in line with the best international practices.
The Executive Board of Sberbank approved the principles of ethics for the development and application of artificial intelligence (AI) technologies in the Sber Group.
Principle #1. Controllability and manageability of AI systems (Secure AI)
• The development and application of AI technologies at Sber are safe, manageable and controllable to the fullest extent possible.
• Sberbank takes into account the possible risks associated with the safety of AI technologies, and does not allow the AI systems to get out of control or cause harm to humans.
• The introduction of AI at Sber is never an end in itself, and the technology in operation must enjoy the confidence of clients, employees and society.
Principle #2. Transparency and predictability of AI technology functioning (Explainable AI)
• Sber is responsible for the use of AI systems in its activities. Sber strives to ensure maximum transparency, internal control and predictability of the process and results of their work.
• AI technologies are applied by Sber in compliance with the laws, including the confidentiality requirements, and with respect to human privacy and trade secrets. When working with AI technologies, data are kept safe, and AI systems are never used for illegal processing of personal information of citizens or information of Sber’s corporate clients.
Principle #3. Stability and reliability of AI systems (Reliable AI)
• Sber ensures the stable and reliable functioning of AI systems.
• Sber maintains the proper level of technical facilities and creates the most favored conditions for the development and implementation of reliable AI systems at Sber.
• Sber provides the highest level guarantees of all human rights and freedoms when using AI, and supports only the lawful use of AI technologies.
Principle #4. Responsible application of AI (Responsible AI)
• When implementing AI, Sber's focus is always on the needs of clients and employees, while AI technologies should be used to improve customer experience.
• Sber takes into account and responsibly treats all concerns that arise in connection with the use of AI technologies.
Principle # 5. Unbiased AI (Fair AI)
• AI technologies are used by Sber fairly and objectively and on a level playing field.
• Sber is committed to ensuring that AI technologies are beneficial for the human development.
To implement these principles, a working group on the AI technology development and application ethics was formed (as a subgroup of the ESG committee), which:
• addresses disputed ethical issues related to the development and application of AI
• facilitates the formation of Sber employees’ corporate ethics in the field of AI
• holds responsibility for the implementation of the principles at Sber
• at the request of the internal customer, examines the risks of using AI technologies in the work of Sber and evaluates the impact of their implementation jointly with the Sber Group Risk Committee
• monitors the effectiveness of the implementation of the principles and their perception in society.
These principles are mandatory for all members of the Sberbank Group and developed based on the requirements of the National Strategy for the Development of Artificial Intelligence for the period until 2030, approved by the President of the Russian Federation. Adherence to these principles will help ensure safety and increase the confidence of society and Sber's customers in AI, which is already used in most of the company's processes, products and services.
“Artificial intelligence technologies, like any others, bring not only opportunities, but also risks. By approving the ethical principles of the development and application of AI technologies and creating mechanisms for their implementation, we act in line with the global best practices and in compliance with the National Strategy for the AI Development. This is a large-scale and complex issue, so our rules and principles are to be further improved and elaborated on. We are open to the opinion of the expert community, which we will take into account when updating our principles. We have not yet found answers for a number of ethical problems arising from the use of AI, even taking into account the best global practices we studied in this area, so our principles will be undoubtedly improved as humanity finds appropriate solutions. I want to note particularly that those are principles of Sber and we do not impose them on anyone beyond our group. Moreover, we do not consider these principles perfect, and other companies do not have to use them in their activities”, says Herman Gref, CEO and Chairman of the Executive Board of Sberbank