AI Index: State of AI in 13 Charts

"

""The index is an independent initiative at the Stanford Institute for Human-Centered Artificial Intelligence (HAI), led by the AI Index Steering Committee, an interdisciplinary group of experts from across academia and industry.

This year’s report covers the rise of multimodal foundation models, major cash investments into generative AI, new performance benchmarks, shifting global opinions, and new major regulations.

A Move Toward Open-Sourced

This past year, organizations released 149 foundation models, more than double the number released in 2022. Of these newly released models, 65.7% were open-source (meaning they can be freely used and modified by anyone), compared with only 44.4% in 2022 and 33.3% in 2021.

bar chart showing that closed models outperformed open models across tasks

But At a Cost of Performance?

Closed-source models still outperform their open-sourced counterparts. On 10 selected benchmarks, closed models achieved a median performance advantage of 24.2%, with differences ranging from as little as 4.0% on mathematical tasks like GSM8K to as much as 317.7% on agentic tasks like AgentBench.

Bar chart showing Google has more foundation models than any other company

Biggest Players

Industry dominates AI, especially in building and releasing foundation models. This past year Google edged out other industry players in releasing the most models, including Gemini and RT-2. In fact, since 2019, Google has led in releasing the most foundation models, with a total of 40, followed by OpenAI with 20. Academia trails industry: This past year, UC Berkeley released three models and Stanford two.

Line chart showing industry far outpaces academia and government in creating foundation models over the decade

Industry Dwarfs All

If you needed more striking evidence that corporate AI is the only player in the room right now, this should do it. In 2023, industry accounted for 72% of all new foundation models.

Chart showing the growing costs of training AI models

Prices Skyrocket

One of the reasons academia and government have been edged out of the AI race: the exponential increase in cost of training these giant models. Google’s Gemini Ultra cost an estimated $191 million worth of compute to train, while OpenAI’s GPT-4 cost an estimated$78 million. In comparison, in 2017, the original Transformer model, which introduced the architecture that underpins virtually every modern LLM, cost around $900.

Bar chart showing the united states produces by far the largest number of foundation models

What AI Race? At least in terms of notable machine learning models, the United States vastly outpaced other countries in 2023, developing a total of 61 models in 2023. Since 2019, the U.S. has consistently led in originating the majority of notable models, followed by China and the UK.

Line chart showing that across many intellectual task categories, AI has exceeded human performance

Move Over, Human

As of 2023, AI has hit human-level performance on many significant AI benchmarks, from those testing reading comprehension to visual reasoning. Still, it falls just short on some benchmarks like competition-level math. Because AI has been blasting past so many standard benchmarks, AI scholars have had to create new and more difficult challenges. This year’s index also tracked several of these new benchmarks, including those for tasks in coding, advanced reasoning, and agentic behavior.

Bar chart showing a dip in overall private investment in AI, but a surge in generative AI investment

Private Investment Drops (But We See You, GenAI)

While AI private investment has steadily dropped since 2021, generative AI is gaining steam. In 2023, the sector attracted $25.2 billion, nearly ninefold the investment of 2022 and about 30 times the amount from 2019 (call it the ChatGPT effect). Generative AI accounted for over a quarter of all AI-related private investments in 2023.

Bar chart showing the united states overwhelming dwarfs other countries in private investment in AI

U.S. Wins $$ Race

And again, in 2023 the United States dominates in AI private investment. In 2023, the $67.2 billion invested in the U.S. was roughly 8.7 times greater than the amount invested in the next highest country, China, and 17.8 times the amount invested in the United Kingdom. That lineup looks the same when zooming out: Cumulatively since 2013, the United States leads investments at$335.2 billion, followed by China with $103.7 billion, and the United Kingdom at$22.3 billion.

Infographic showing 26% of businesses use AI for contact-center automation, and 23% use it for personalization

Where is Corporate Adoption?

More companies are implementing AI in some part of their business: In surveys, 55% of organizations said they were using AI in 2023, up from 50% in 2022 and 20% in 2017. Businesses report using AI to automate contact centers, personalize content, and acquire new customers.

Bar chart showing 57% of people believe AI will change how they do their job in 5 years, and 36% believe AI will replace their jobs.

Younger and Wealthier People Worry About Jobs

Globally, most people expect AI to change their jobs, and more than a third expect AI to replace them. Younger generations — Gen Z and millennials — anticipate more substantial effects from AI compared with older generations like Gen X and baby boomers. Specifically, 66% of Gen Z compared with 46% of boomer respondents believe AI will significantly affect their current jobs. Meanwhile, individuals with higher incomes, more education, and decision-making roles foresee AI having a great impact on their employment.

Bar chart depicting the countries most nervous about AI; Australia at 69%, Great Britain at 65%, and Canada at 63% top the list

While the Commonwealth Worries About AI Products

When asked in a survey about whether AI products and services make you nervous, 69% of Aussies and 65% of Brits said yes. Japan is the least worried about their AI products at 23%.

Line graph showing uptick in AI regulation in the united states since 2016; 25 policies passed in 2023

Regulation Rallies

More American regulatory agencies are passing regulations to protect citizens and govern the use of AI tools and data. For example, the Copyright Office and the Library of Congress passed copyright registration guidance concerning works that contained material generated by AI, while the Securities and Exchange Commission developed a cybersecurity risk management strategy, governance, and incident disclosure plan. The agencies to pass the most regulation were the Executive Office of the President and the Commerce Department. """

Author

ABN ASIA was founded by people with deep roots in academia, with work experience in the US, Holland, Hungary, Japan, South Korea, Singapore, and Vietnam. ABN Asia is where academy and technology meet opportunity. With our cutting-edge solutions and competent software development services, we're helping businesses level up and take on the global scene. Our commitment: Faster. Better. More reliable. In most cases: Cheaper as well.

Feel free to reach out to us whenever you require IT services, digital consulting, off-the-shelf software solutions, or if you'd like to send us requests for proposals (RFPs). You can contact us at contact@abnasia.org. We're ready to assist you with all your technology needs.

© ABN ASIA