(Report) AI Index Report 2023: Industry Takes Over the Lead in AI Development and Deployment
The annual AI Index Report has highlighted a significant shift in the world of AI.
The annual AI Index Report has highlighted a significant shift in the world of AI, with industry players now dominating the development and deployment of AI applications, leaving academia and government behind.
The report, compiled by researchers from Stanford University and leading AI companies, suggests that the AI industry is entering a new phase of development, marked by the mainstream adoption of various AI tools. However, as cutting-edge AI requires vast resources, including data, staff, and computing power, academia is struggling to keep up with industry in terms of producing state-of-the-art AI systems.
The report emphasizes that decisions about deploying this technology and managing risks and opportunities now lie firmly in the hands of corporate players.
Selected highlights
Private investment in AI has experienced its first decline in a decade
Despite a steady increase in global private investment in AI over the years, the figure dropped by 26.7 percent from 2021 to $91.9 billion in 2022.
The environmental costs of training large AI models are becoming increasingly evident
In 2022, a study revealed that the carbon emissions from training the BLOOM AI language model were 25 times higher than that of flying one passenger from New York to San Francisco and back, whereas OpenAI's GPT-3 had a carbon cost 20 times that of BLOOM.
On a positive note, AI has the potential to mitigate carbon emissions. For instance, in 2022, DeepMind, a subsidiary of Google, developed an AI system named BCOOLER, which reduced energy consumption by 12.7 percent in a three-month trial in the company's data centers by optimizing cooling procedures. However, it remains uncertain if Google has implemented this system more widely.
Chinese people are more optimistic about AI than Americans
Interestingly, an Ipsos survey conducted in 2022 revealed that about 78 percent of Chinese respondents believed that "products and services using AI have more benefits than drawbacks," making them the most enthusiastic group about AI, followed by Saudi Arabians (76 percent) and Indians (71 percent). In contrast, only 35 percent of American respondents agreed with the statement.
The instances of AI misuse are rapidly increasing
AIAAIC, a database that monitors cases related to the unethical use of AI, has reported a 26-fold surge in the number of AI incidents and controversies since 2012. This increase is a clear indication of the rising utilization of AI technologies and heightened awareness of potential misuse.