上海的陆家嘴

最近,OpenAI推出的ChatGPT等生成式AI模型引起广泛关注。这些模型需要训练极大量的参数,因此训练过程消耗了大量计算资源,进而产生了大量碳排放。据报道,训练GPT-3这一模型就消耗了几百万美元的计算成本。鉴于生成式AI的环境影响日益凸显,美国民主党人已提出相关立法,要求AI公司报告碳排放和能源消耗数据。该举措有助于监管AI行业的外部成本。

Title: Huge energy consumption behind generative AI
Keywords: AI model training, power consumption, legislative oversight

News content: Recently, generative AI models like ChatGPT from OpenAI have attracted wide attention. These models need to be trained on an extremely large number of parameters, thus consuming huge computational resources and generating massive carbon emissions in the training process. It is reported that training GPT-3 alone has cost millions of dollars in computing. In light of the increasing environmental impact of generative AI, the US Democratic Party has proposed related legislation requiring AI companies to report data on carbon emissions and energy consumption. This move will help regulate the external costs of the AI industry.

【来源】https://www.thepaper.cn/newsDetail_forward_26438846

Views: 1

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注