Chinese local governments have been told they may use artificial intelligence to help decision-making, but must ensure they do not leak any sensitive data or “blindly” pursue technological superiority.
Advertisement
New guidelines, published by the Cyberspace Administration of China and the National Development and Reform Commission on Friday, said AI models could make government more efficient and benefit the public.
It listed areas where AI models could be used. These included online consultations to answer questions from the public; the use of drones and surveillance cameras to spot decaying infrastructure; and to help draft documents, carry out research and fill out forms.
Big data may be used to evaluate policy, provide natural disaster alerts and generate emergency plans, the document said.
But the guidelines warned that AI models needed to be chosen cautiously. Local governments were told they must protect data and individual information when training AI models, regulate the use of these models and make sure they were not given access to state secrets.
Advertisement
There must be a complete safety evaluation before these models go online, the guidelines also stated, and government workers should be trained in the use, theory and technology of AI as well as safety and ethics.
The guidelines were issued at a time when the central authorities are looking to improve government efficiency by reducing the burden on grass-roots officials and cutting back on “formalism” – a term referring to the need to attend large numbers of meetings and produce extensive reports.