The Serious Obligation Facing Digital Companies in the Age of AI

Almost before we realized it, artificial intelligence (AI) has found its way into our everyday lives, and is already performing a diverse range of functions. The Ricoh Group, also, has applied one of the latest cutting-edge AI technologies--deep learning functionality--to the built-in chip inside PENTAX K-1, Single-lens reflex camera, a top-of-the-line model, dramatically increasing accuracy with regard to aspects such as photometry and exposure.

Specifically, in addition to face recognition functionality and image sensors that instantly adjust brightness and photographic sensitivity, the camera is also equipped with capabilities such as scene analysis, to distinguish whether the scene being shot is a motionless natural landscape, or a speed-filled composition with sporting athletes. Thanks to these features, even amateur photographers can now enjoy photographic technology comparable with that used by professionals.

Meanwhile, interest is gathering as to the extent to which we can avoid social issues caused by malfunctioning AI technologies. It is true that if the computer algorithms that make AIs work were to run out of control they would have an impact on society and our lives in general. In the worst case scenario, human lives could even be lost. If a camera is faulty, we can simply throw bad pictures away, but the same cannot be said in the result caused by a malfunctioning automated car.

Amidst these concerns, the European Union (EU) has introduced the General Data Protection Regulation (GDPR), and started initiatives with a view to ensuring the protection and appropriate handling of data. Japan, also, is in the process of organizing basic regulation for the utilization of AI. However, these initiatives have only just begun. It goes without saying that when it comes to these algorithms, it is essential for tests to be conducted in advance at the company level, where the development and production of these technologies take place, to make doubly sure that they function appropriately as they should.

It is also important to make use of insurance functionality to prepare for the unlikely event of accidents caused by such technologies. If we create schemes that offer insurance and salvation from damage, based on the assumption of the risk of AIs spiraling out of control, then users will have greater peace of mind. It would also be meaningful for the organizations that provide these kinds of insurance to take a central role in establishing neutral institutions to check the content of algorithms used in such systems.

Whatever the case may be, the fusion of existing sophisticated analog technologies with cutting-edge AI functionality will lead to the provision of more convenient goods and services. For digital companies such as Ricoh, the extent to which issues accompanying AI-related malfunctions can be averted will become a serious obligation, and a problem which must be overcome in the future.