The latest 2022 AI Index highlights an AI investment boom, and focus on ethics
The field of artificial intelligence (AI) is at a critical crossroad, according to the 2022 AI Index, an annual study of AI impact and progress at the Stanford Institute for Human-Centred Artificial Intelligence (HAI) led by an independent and interdisciplinary group of experts from across academia and industry. Last year saw the globalisation and industrialisation of AI intensify, while the ethical and regulatory issues of these technologies multiplied.
The 2022 AI Index is one of the most comprehensive reports about AI to date. It measures and evaluates the rapid rate of AI advancement via a cross-sector lens, from research and development to technical performance and ethics, AI policy and governance, the economy and education, and more.
Artificial intelligence is getting cheaper, better at the tasks we assign it, and more widespread — but concerns over bias, ethics, and regulatory oversight remain.
Here are some of the highlights.
Computers that analyse images and understand speech
Last year, the research community was focused on applying AI to computer vision, a subfield that teaches machines to understand images and videos in order to get good at classifying images, recognising objects, mapping the position and movement of human body joints, and detecting faces.
For image classification, the most popular database used to train AI models is called ImageNet. Some researchers pre-train their models on additional datasets before exposing them to ImageNet. But models still make mistakes, on average misidentifying one out of ten images. The model that performs the best is from the Google Brain Team. In addition to identifying images and faces, AI can also generate fake images that are nearly indistinguishable from real ones, and to combat this, researchers have been working on deepfake detection algorithms that are based on datasets like FaceForensics++.
Natural language processing is slowly making progress in English language understanding, summarising, inferring reasonable outcomes, identifying emotional context, speech recognition and transcription, and translation. For basic reading comprehension, AI can perform better than humans, but humans still have an edge when it comes to interpreting context clues. On the other hand, AI ethicists are worried that bias could affect large language models that draw from a mixed bag of training data.
Meanwhile, tech companies like Amazon, Netflix, Spotify, and YouTube have been improving the AI used in recommendation systems. The same is true for AI’s role in reinforcement learning, which has enabled it to react and perform well in virtual games such as chess. Reinforcement learning can also be used to teach autonomous vehicles tasks like changing lanes, or help data models predict future events.
As AI appears to have become better at doing what we want it to do, the cost to train it has come down as well, dropping by over 60 per cent since 2018. Meanwhile, a system that would’ve taken six minutes to train in 2018 would now only take a little over 13 seconds. Accounting for hardware costs, in 2021, an image classification system would take less than $5 to train, whereas that cost would’ve been over $1,000 in 2017.
The new report shows several key advances in AI in 2021:
- Private investment in AI has more than doubled since 2020, in part due to larger funding rounds. In 2020, there were four funding rounds worth $500 million or more; in 2021, there were 15.
- AI has become more affordable and higher performing. The cost to train an image classification has decreased by 63.6 per cent, and training times have improved by 94.4 per cent since 2018. The median price of robotic arms has also decreased 46.2 per cent in the past five years.
- The US and China have dominated cross-country research collaborations on AI as the total number of AI publications continues to grow. The two had the greatest number of cross-country collaborations in AI papers in the last decade, producing 2.7 times more joint papers in 2021 than between the United Kingdom and China — the second highest on the list.
- The number of AI patents filed has soared—more than 30 times higher than in 2015, showing a compound annual growth rate of 76.9 per cent.
At the same time, the report also highlights growing research and concerns on ethical issues as well as regulatory interests associated with AI in 2021:
- Large language and multimodal language-vision models are excelling on technical benchmarks, but just as their performance increases, so do their ethical issues, like the generation of toxic text.
- Research on fairness and transparency in AI has exploded since 2014, with a fivefold increase in publications on related topics over the past four years.
- Industry has increased its involvement in AI ethics, with 71 per cent more publications affiliated with industry at top conferences from 2018 to 2021.
- Globally, AI regulation continues to expand. Since 2015, 18 times more bills related to AI were passed into law in legislatures of 25 countries around the world and mentions of AI in legislative proceedings also grew 7.7 times in the past six years.
