The state of AI.

Stanford just dropped this 386-page report on the state of AI.

I created a TL;DR of the 10 trends to save you time.

1. Industry races ahead of academia.
State-of-the-art AI systems hungry for data, compute, and money.
Until 2014, academia built the best ML models.
In 2022, the industry produced 32 ML models vs 3 by academia.

2. Misuse of AI is rapidly rising.
AI incidents and controversies increased 26 times since 2012.
E.g Deepfake video of Ukrainian President Zelenskyy surrendering.

3. AI is both helping and harming the environment.
AI models are compute and energy hungry. New models show promise.

4. AI’s the world’s best new scientist.
In 2022 AI helped hydrogen fusion, improve the efficiency of matrix manipulation, and generate new antibodies.

5. Policymaker interest in AI is on the rise. 127 countries now have bills containing “artificial intelligence”.

6. For the first time in the last decade, year-over-year private investment in AI decreased.
$91.9 billion in 2022(26.7% decrease). Still, in 2022 private investment was 18 times > 2013.

7. Performance saturation on traditional benchmarks.
AI improved, but YoY improvement on many benchmarks are marginal. New benchmarks coming up.

8.Companies adopting AI has plateaued(at 50%- 60%). But those who adopt AI saw cost decrease and revenue increase.

9. Chinese citizens feel positive about AI products and services. Americans, not so much.

10. The demand for AI skills is increasing across every sector.