The 2023 AI Index Report, an initiative led by the Stanford Institute for Human-Centered Artificial Intelligence (HAI), provides a comprehensive overview of trends and developments in the field of artificial intelligence. The report aims to offer valuable data and insights to decision-makers seeking to advance AI in a responsible and ethical manner.
Key Takeaways from the 2023 AI Index Report
Industry races ahead of academia:
The development of significant machine learning models has shifted from academia to industry. In 2022, industry produced 32 significant machine learning models compared to just three from academia. This is attributed to industry’s greater access to data, computing power, and funding.
Performance saturation on traditional benchmarks:
Although AI continues to achieve state-of-the-art results, year-over-year improvement on many benchmarks is marginal. New benchmarking suites such as BIG-bench and HELM are being introduced to address this saturation.
AI’s environmental impact:
AI systems can have significant environmental impacts, with some models emitting substantial amounts of carbon. However, AI can also contribute positively by optimizing energy usage, as demonstrated by reinforcement learning models like BCOOLER.
AI accelerates scientific progress:
AI models are contributing to advancements in science, aiding in areas such as hydrogen fusion, matrix manipulation, and antibody generation.
Rise in AI misuse incidents:
Incidents of ethical misuse of AI have increased, with 2022 seeing deepfake videos and the use of call-monitoring technology in prisons. This highlights the importance of ethical considerations in AI deployment.
Increasing demand for AI-related skills:
Across nearly all industrial sectors in the United States, there is an increasing demand for workers with AI-related skills, as evidenced by the rise in AI-related job postings.
Decrease in private investment in AI:
For the first time in a decade, year-over-year private investment in AI decreased in 2022, with global AI private investment dropping by 26.7%. However, AI investment has still significantly increased over the last decade as a whole.
AI adoption plateaus, but adopters pull ahead:
AI adoption has plateaued, with 50% to 60% of companies adopting AI. However, those companies that have adopted AI report meaningful cost decreases and revenue increases.
Policymaker interest in AI increases:
The number of AI-related bills passed into law and mentions of AI in legislative proceedings have both increased, indicating growing interest from policymakers.
Varied public perception of AI:
Chinese citizens have a positive perception of AI products and services, with 78% agreeing they have more benefits than drawbacks. In contrast, only 35% of Americans share this view.
- Chapter 1: Research and Development
- Chapter 2: Technical Performance
- Chapter 3: Technical AI Ethics
- Chapter 4: The Economy
- Chapter 5: Education
- Chapter 6: Policy and Governance
- Chapter 7: Diversity
- Chapter 8: Public Opinion
Read the full 2023 AI Index Report