Magazine #1 | Summer 2022
Sustainable AI:
Facts & Figures
Stanford University’s annual AI Index Report tracks, collates, distills and visualizes data related to Artificial Intelligence. Its aim is to provide a more thorough and nuanced understanding of the field of AI. We present some of the data from this year’s edition that is relevant from a sustainability perspective.
Growth
Private AI Investment
The AI industry is growing rapidly, especially investments in data management, processing and clouds. In 2021, they increased by more than two and a half times compared to the previous year and amounted to around USD 4.69 billion. Two of the four largest private investments in 2021 went to data management companies.
Governance:
Federal AI Legislation in the United States
The federal legislative record in the United States shows a sharp increase in the total number of proposed bills that relate to AI. In 2015, just one federal bill was proposed, while in 2021, there were 130. The number of bills related to AI being passed has not kept pace with the growing volume of proposed AI-related bills. This gap was most evident in 2021, when only 2 percent of all AI-related bills were ultimately passed into law.
Pre-Trained:
Number of Commercially Available MT Systems
The growing interest in machine translation is reflected in the rise of commercial machine translation services such as Google Translate. Since 2017, there has been a nearly fivefold increase in the number of commercial machine translators on the market – but only few pre-trained models. 2021 saw the introduction of the open-source MT services M2M-100, mBART and OPUS.
Diversity:
Women In Machine Learning
Diversity in AI is key against discrimination. Founded in 2006, Women in Machine Learning (WiML) is an organization dedicated to supporting and increasing the impact of women in Machine Learning. This data illustrates the number of attendees at WiML workshops over the years at NeurIPS – one of the most important AI and ML conferences.
Risks:
Consideration and Mitigation of Risks from Adopting AI
Risks not only need to be recognized, but also to be actively addressed. Currently, gaps remain between recognizing risks and acting upon them — a gap of 10 percentage points regarding risks relating to equity and fairness (29 percent to 19 percent), 12 percentage points for regulatory compliance (48 percent to 36 percent), 13 for personal/individual privacy (41 percent to 28 percent), and 14 for explainability (41 percent to 27%).
Growth:
AI Patents
AI is becoming more pervasive: The number of patents filed in 2021 is more than 30 times higher than in 2015, showing a compound annual growth rate of 76.9 percent.
Governance:
Global Legislation Records on AI
Governments and legislative bodies across the globe are increasingly seeking to pass laws to regulate the development of AI. An analysis of laws passed in 25 countries by their legislative bodies that contain the words “Artificial Intelligence” showed that, taken together, a total of 55 AI-related bills have been passed.
Growth:
Number of Newly Funded AI Companies
Cultural context matters in AI development. The gap between the two leading countries and the rest of the world in AI development is significant – there is a heavy dominance of US based AI development, followed by China and more distantly by select European and Asian countries.
