TOKYO -- Sony will start screening its products that use artificial intelligence for ethical risks as early as spring 2021, Nikkei has learned. If a product is deemed ethically inappropriate, the company will either improve the product or halt its development.
While AI offers consumers more convenience, it could unwittingly cause embarrassing problems, such as revealing social biases due to a lack of understanding as to what and how AI is learning.
As AI finds its way into more devices, the responsibilities of developers are increasing, and companies are strengthening guidelines governing AI ethics.
Sony plans to reflect its guidelines for AI ethics into existing quality control procedures. Japanese companies rarely consider AI ethics when measuring product safety.
From product development to post-launch, Sony will examine AI-imbued items as per the guidelines, including privacy protections. Any products not meeting the criteria for ethical AI performance will be modified or dropped.
Sony's CEO appoints the members of its AI Ethics Committee from existing staff, along with its chairperson. The committee examines the degree of compliance as defined by the guidelines.
Even products well into development are subject to being dropped. Products that have already been sold can also be recalled if problems are discovered.
There is growing global concern over the risks of discrimination and abuse posed by AI. In 2019, a U.S. regulator investigated Goldman Sachs for possible sex discrimination in the way it sets credit card limits. This followed a 2018 incident at Google, whose employees discovered that the company's AI technology was being used in unmanned military aerial vehicles, prompting thousands of staff to protest to top management.
The Link LonkDecember 22, 2020 at 04:09PM
https://ift.tt/34Bc6YC
Sony to review ethics of AI products to avoid social bias - Nikkei Asia
https://ift.tt/2ZeUDD8
Sony
No comments:
Post a Comment