“Machine learning through the use of fairness to keep that”discussion is being carried out. Machine learning technology researchers in the Community [Society]”of the Japan Society of artificial intelligence the Ethics Committee””Japan Software Science Society, machine learning in engineering society [MLSE]””ieice information-theoretic learning theory and machine learning research”, in machine learning by the use of the fairness issue to the public at large to tell,12 November 10 issued a statement on that.
In a statement, the following 2 points are important as you, and the content excerpt to.
■The society’s statement”machine learning and fairness concerning the statement” machine learning tools just in human decision-making to assist this.
“Machine learning is based on past cases predicts the future, from a biased based in the past to predict the future,and biased ones that don’t. Also, in the past, and different”should be the future”if you are looking for a machine learning prediction and judgment is fairness to the lack of a like a human learning to carefully needs to intervene.”  our machine learning in fairness contributes.
“Machine learning by fairness to this can be to prevent the problem, as well as machine learning as fairness the way to define, discuss and sincere measures.”
■AI overconfidence and
AI is the”past data”to analyze reasoning have a data bias, if there is, of course, to the conclusion is biased as. For example, recently in the recruitment of spot starting to be used, the AI of the conclusion of reason is not shown, such as a sense of injustice was out.
Recruitment in the field of AI the conclusion of the reference, the Examiner’s biased, prevents such over the course of several days. AI is based on historical data”is used, a job likely to be”such as to draw conclusions and to adopt, that is, AI is the conclusion a description of the need and been told.
Or in-house personnel, AI is the data of the past from the placement change, if the past data is what kind of environment was the conclusion of whether there are many important questions,”unconventional talent”buried no other bad effects happen.
These considerations have,in the past HR professionals of forethought is required,at that time, the corporate side of the business needs are legitimate, and the range personnel are that need to be done.
The computer began to be used at the time and in the same way,the computer of out the conclusion to the”deification”can be hazardous. AI also human education and experience, in the same way as the environment[the data]will depend on the scope is greater still.
“Machine learning artificial intelligence is not?” And constantly question, and the forethought that is required of you. “People make mistakes, the AI are all committed to get”and thinking,”all matters of relevance to structural thinking”you have come. [Article: kenzoogata・The article list to look at]