By Kantaro Komiya and Sam Nussey
TOKYO (Reuters) – Japan’s privacy watchdog said on Friday it has warned OpenAI, the Microsoft-backed startup behind the ChatGPT chatbot, not to collect sensitive data without people’s permission.
OpenAI should minimise the sensitive data it collects for machine learning, the Personal Information Protection Commission said in a statement, adding it may take further action if it has more concerns.
Regulators around the world are scrambling to draw up rules governing the use of generative artificial intelligence (AI), which can create text and images, the impact of which proponents compare to the arrival of the internet.
While Japan has been on the backfoot with some recent technology trends, it is seen as having greater incentive to keep pace with advances in AI and robotics to maintain productivity as its population shrinks.
The watchdog noted the need to balance privacy concerns with the potential benefits of generative AI including in accelerating innovation and dealing with problems such as climate change.
Japan is the third-largest source of traffic to OpenAI’s website, according to analytics firm Similarweb.
OpenAI CEO Sam Altman in April met Prime Minister Fumio Kishida with an eye to expansion in Japan, ahead of the Group of Seven (G7) leaders summit where Kishida led a discussion on regulating AI.
The EU, a global trendsetter on tech regulation, set up a taskforce on ChatGPT and is working on what could be the first set of rules to govern AI.
In the meantime, the rapid spread of such chatbots has meant regulators have had to rely on existing rules to bridge the gap.
Italian regulator Garante had ChatGPT taken offline before the company agreed to install age verification features and let European users block their information from being used to train the system.
Altman last week said OpenAI had no plans to leave Europe after earlier suggesting the startup might do so if EU regulations were too difficult to comply with.
