Shulou(Shulou.com)11/24 Report--

CTOnews.com, July 13 (Xinhua) the Global Technology Policy Committee on Computing machines (ACM TPC) of the American Association for generative artificial Intelligence (Association for Generative AI) recently issued the principles for the Development, deployment and use of generative artificial Intelligence Technology, which has been adopted by the US Technology Policy Committee (USTPC) and the European Technology Policy Committee (Europe TPC) of ACM.

It is understood that the principle aims to promote fair, accurate and useful decision-making on generation and all other artificial intelligence technologies. However, ACM did not urge the implementation of specific legislation in this regard.

Jeremy Epstein, chairman of the US Technology Policy Committee of the ACM, said, "We will respond directly to the views of the legislature, but will not publicly declare our position on the bill or some of its contents."

Founded in 1947, ACM is known as "the world's largest association for scientific and educational computing", with about 110000 students and professional members. They are of the view that the growing power of generative AI may pose a serious risk to society, and that the current guidelines for the safe use of the technology are insufficient, so new standards need to be established to ensure that the possible hazards of the technology can be avoided.

Among the principles released by the American artificial Intelligence Association, four principles are specific to generative AI, and the other four principles are adapted from TPC's "responsible algorithm system statement" in 2022. CTOnews.com puts the full text of the principles below:

Restrictions and guidelines for the deployment and use of specific principles of generative AI: written or revised laws and regulations shall be reviewed and applied in consultation with all stakeholders to limit the deployment and use of generative artificial intelligence technologies as needed to minimize harm. In the absence of clear and adequate safeguards, including "human in the loop" and a clear consensus among relevant stakeholders that the benefits of the system will significantly outweigh its potential negative impact, no high-risk artificial intelligence system should be allowed to operate. One way is to define a risk level, with the highest level of unacceptable risk and the lowest level of minimum risk.

Ownership: intellectual property (IP) laws and regulations have not fully taken into account the inherent aspects of the structure and function of generative artificial intelligence systems.

Personal data control: generative artificial intelligence systems should allow people to choose not to use their data to train the system or facilitate its information generation.

Correctability: the provider that generates the artificial intelligence system should create and maintain a common repository that can record errors made by the system and can choose to correct them.

Transparency of the previous principle of adaptation: any application or system that uses generative artificial intelligence should be clearly disclosed to the appropriate stakeholders.

Auditability and questionability: providers that generate artificial intelligence systems should ensure that system models, algorithms, data and outputs are recorded as much as possible (with due regard to privacy) so that they can be audited and / or challenged where appropriate.

Limiting environmental impacts: in view of the significant impact of generative artificial intelligence models on the environment, it is recommended to reach a consensus on methodology to measure, attribute and actively reduce such impacts.

Higher security and privacy: generative artificial intelligence systems are vulnerable to a variety of new security and privacy risks, including new attack vectors and malicious data disclosure.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.