The rise of artificial intelligence is not slowing down yet, with new figures showing a 34.5 percent increase in publishing AI research from 2019 to 2020. That’s higher percentage growth than 2018 to 2019. , when the volume of publications increased by 19.6 percent.
China continues to be a growing force in AI R&D, outpacing the US in general magazine citations in artificial intelligence research last year. The country already publishes more articles on AI than any other country, but the United States still has more articles cited at AI conferences, an indicator of the novelty and importance of the underlying research.
These figures come from the fourth annual AI Index, a collection of statistics, benchmarks and milestones intended to measure global progress in artificial intelligence. The report was compiled with the help of Stanford University and you can read the 222 pages here.
In many ways, the report confirms trends identified in recent years: the sheer volume of AI research is growing across a number of metrics, China continues to be increasingly influential, and investors are pumping even more money into AI companies. .
However, the details reveal subtleties about the AI scene. For example, while private investment in AI increased 9.3 percent in 2020 (a larger increase than from 2018 to 2019 of 5.7 percent), the number of newly funded companies that received funds decreased for a third year. consecutive. There are several ways to interpret this, but it suggests that investors expect the winner-take-all dynamic that has defined the tech industry, in which digital economies of scale tend to reward a few dominant players, will be replicated in the world of AI. .
The section of the report on technical advances also confirms the main trends in AI capabilities, the most important of which is the industrialization of computer vision. This field has seen incredible progress during the rise of artificial intelligence, with services like facial and object recognition now common. Similarly, generative technologies, which can create video, images, and audio, continue to increase in quality and availability. As the report notes, this trend “promises to generate a huge variety of downstream applications of AI for socially useful and less useful purposes.” Useful apps include cheaper computer-generated media, while malicious results include misinformation and AI revenge porn.
One area of AI research that seems to be just beginning to make sense is biotechnology. The drug discovery and design sector received the largest private investment of any sector in 2020 ($ 13.8 billion, 4.5 times more than in 2019), and experts surveyed for the AI Index report cited DeepMind’s AlphaFold program, which uses machine learning to fold proteins, like one. of the biggest breakthroughs in AI in 2020 (the other frequently cited breakthrough last year was OpenAI’s GPT-3 text generation program).
However, one area where the Index AI report struggles to measure progress is in ethics. This is a wide-ranging area, encompassing everything from facial recognition politics to algorithmic bias, and the discussion of these topics is becoming increasingly prominent. In 2020, stories like Google’s firing of researcher Timnit Gebru and IBM’s exit from the facial recognition business fueled discussions about how artificial intelligence technology should be applied. But while companies are happy to speak lip-service about ethical principles, the report notes that most of these “commitments” are non-binding and lack institutional frameworks. As noted in the past: the AI ethic for many companies is simply a way to curb criticism.