DeepMind establishes "DeepMind Ethics & Society (DMES)" to study "ethics of AI"

Deep Mind under the umbrella of Google is a research department for solving various ethical problems that artificial intelligence (AI) technology has "DeepMind Ethics & Society (DMES)"Was established. It is judged that it is indispensable to solve ethical problems for future evolution of AI and utilization in society.

Deep Mind Ethics & Society | Deep Mind
https://deepmind.com/applied/deepmind-ethics-society/

AI who broke down the Go champion "AlphaGoDeep Mind, known for its development, launched the DMES department to study issues related to AI's ethics. In DMES, six researchers including Professor AI development expert as well as prominent philosopher of Oxford University Professor Nick Obidstrom participated in the "DeepMind Research Fellows" position, discussion on AI technology and ethics I will do research.

The famous ethical problem concerning AI is "Is the automatic driving car to prioritize any human life in emergency?" Is it permissible to kill a pedestrian to protect the driver's life in the car? Or should we accept the death of the driver to save a lot of people? It is an argument asking the ultimate choice.

"Do you sacrifice the driver to help passers-by?" How does the fully automatic driving car correspond to the moral dilemma that relates to human's lives? - GIGAZINE


There is another argument that guaranteeing human rights may be impaired by introducing AI into the criminal justice system. For example, in the United States, in the criminal justice system, use AI technology to calculate the escape risk of a criminal defendant, make it a material for decision of parole release, and list a person with high risk of becoming a perpetrator of gun crime doing. However, as it is known that bias (bias) will occur in the judgment of AI, there is concern as to whether human rights abuses will occur by leaving judgment of criminal justice to AI.

AI learns female discrimination and racial discrimination from human language - GIGAZINE


Everyone acknowledges the importance that AI technology plays in economic development, but because AI's influence is so great that AI can create an AI threat theory that it could destroy humanity over its handling . In order to develop AI technology to utilize it as a knowledge of human race sharing, it can be said that the voice calling for the necessity of ethical restrictions is increasing day by day.

Artificial intelligence is potentially dangerous than nuclear weapons, Dr. Hawking warns that "Artificial intelligence exceeds humans within 100 years" - GIGAZINE


While AI threat theory was partially whispered, Deep Mind, who started developing AI with higher versatility, worked on reducing health care costs in cooperation with the UK National Health Service (NHS), but 1.6 million people If the access right to the medical data of the patient is reported, criticism that "there is a problem from the viewpoint of protection of personal information" and "the legal basis concerning the data acquisition method is doubtful" rushes, and the option that the patient can refuse data provision We are faced with problems related to AI and human rights, such as being forced to add to the situation.

Although AI technology with such a sensitive problem, in order to further advance the research, Deep Mind thought that solving various problems concerning AI and ethics can not be avoided is a research result on AI and ethics In addition, DMES was established as an organization independent from the development department in order to collaborate with other researchers to collaborate on this problem.

DeepMind defines its mission as "to make the world better by intelligence", and development of AI technology is one of its powerful methods. Therefore, in AI development, DeepMind has responsibility to open research and investigation. And it is the purpose of the establishment that DMES, which is planning to conduct not only original research but also other researchers with other researchers, provides a basis for discussing the influence of AI on real society. DMES's Bertie Harding co-leader said, "If AI is a useful technology for society, it must be developed to resolve concerns while utilizing it for social priorities." It is.


Deep Mind's DMES already includes Oxford Internet Institute's Digital Ethics Laboratory (Delab), Bath University's Institute for Policy Research (IPR), New York University, British Public Policy Institute (IPPR), Leverwell Center (CFI), UK Research institutions such as the Royal Society such as the United Kingdom and the United States have been named as partners, deepening the discussion on "AI and ethics" and working on research to contribute to the development of AI research. DMES plans to publish the first research paper in the first half of 2018.