Interview for formulation of Sony Group
AI Ethics Guidelines
Focusing on the issues that we can envision arising from the development of AI
Since about 2015, vigorous discussions concerning AI have been taking place globally. While some of these discussions are positive in nature, such as how we can use AI to solve social issues, and how AI can be useful in the development of industry, other arguments are also taking place - Won't AI generate new issues, for example issues of privacy, security, disparity, discrimination? Won't AI replace jobs? Won't AI supplant human thought, or even threaten the existence of humanity? In addition, given that AI learns enormous amounts of data in order to derive answers, we can imagine occasions on which we do not understand the process by which a specific answer was derived, or in which the answer is incomplete due to statistical processing. Based on such risks, guidelines are provided for thinking about what AI should be used for, and about whether or not AI should actually be used in certain cases.
We will incorporate diverse opinions in order to revise and evolve our Guidelines
One of the distinctive features of the Guidelines is that they incorporate ideas that form the foundation of Sony's business. Article 1 of the Guidelines states that we seek to support people in empowering their lives, bringing them emotion, and contributing to the enrichment of our culture. These goals are based on our Founding Prospectus, Sony's origin. From the stage of drafting the Guidelines, we continuously examined their content, not only conducting interviews within the Sony Group, which encompasses businesses including electronics, games, music, movies and finance, but also outside the Group. One of the opinions that we heard was that having the Guidelines might increase the number of necessary business procedures, but it came to be understood that engaging in thinking that incorporated ethical elements from the initial stage of design would actually promote development and contribute to the further development of AI. In the future, we will continue to incorporate diverse opinions and revise our Guidelines.
By promoting dialogue, we will bring diversity into our businesses
Because the Guidelines contain high-level content, the most important thing in concrete examinations of how to apply them in your own works is dialogue. Promoting dialogue is also one of the major goals of the Guidelines. This does not mean dialogue within the company only. Article 2 of the Guidelines, "Stakeholder Engagement," urges us to proactively advance a dialogue with customers and creators, related industries, organizations, academic communities and more. By activating feedback loops with a diverse range of individuals, we will naturally incorporate diversity within our business. In other words, dialogue is also a tool for the promotion of diversity. I believe that the incorporation of diversity is extremely important to the ability to manifest creativity.
AI will expand the potential of every individual, and this will change the world
Whether their focus is science or humanities, a diverse range of people will be involved with AI. Rather than being able to point to a single field as the one you work in, we would like you, say, to possess data science knowledge and actively involve yourself in the development of AI while allowing your fields of specialization to cross-fertilize. This is an opportunity to expand your potential without being tied down by your career up to now. I think that it is more meaningful to use AI to expand the abilities of every individual rather than focusing on how AI will change the society. It is this that will change the world and fill it with emotion. That is part of my dream for AI.