According to the AI Index Report by researchers from Stanford University, AI technology is evolving quickly and it has many effects on the global economy and society. The report shows details about how AI technology is developing, how new AI hardware is bringing changes and what are the costs of AI inference. The report also talked about trends in AI technology and how it is growing in fields like medicine and science. The researchers also emphasized the importance of responsible AI practices.
According to the report, developers have made significant progress in AI technology over the past year and have gotten record-breaking scores on benchmarks like GPQA, MMLU, and SWE-bench which were developed in 2023 to test the limits of AI. Many advancements in AI-generated videos and AI agents have also made many tasks a lot easier. The US has the top-performing AI models but China is catching up as well. In 2024, 40 AI models were published by different organizations in the US as compared to 15 from China and 3 from Europe. Chinese models are showing a lot of progress and are competing with the US models on benchmarks like HumanEval and MMLU.
China is quickly gaining ground in AI because of the efficiency, affordability, and accessibility of its AI systems. A report found that the cost of inference for a system that works like OpenAI’s GPT-3.5 has dropped 280 times over the past two years, and the hardware costs have decreased 30% in the last year. 40% less energy is being used by AI models as compared to a year ago and more powerful models are now being used. Many regions like Southeast Asia, the Middle East, and Latin America are also confident in launching their own powerful LLMs. However, the report also found that even the most powerful AI models are struggling to reach human reasoning skills. AI can do well on tasks like the International Math Olympiad but still faces difficulty when it comes to performing tasks which require logical reasoning.
AI is also continuing to attract billions of dollars of investment, with the US having the largest share. American investors invested $109 billion for AI development, as compared to China’s $9.3 billion and UK’s $4.5 billion in 2024. Generative AI has gotten $33.9 billion in private investment, which is an 18% increase from 2023. The report found that 78% of the global enterprises have deployed AI in their workplaces in 2024, which is a 55% increase from 2023.
AI has also made a big impact on society where it is now being incorporated in transportation and healthcare, with FDA approving 950 AI-powered medical devices. AI vehicles like Waymo’s self-driving taxis are also getting about 150,000 rides per week. However, the public opinion in the US about AI is still cautious, with only 39% of Americans thinking of AI as beneficial as compared to more than 75% of people from Thailand, China, and Indonesia who think of AI as advantageous. On the other hand, there is still a big gap when it comes to incorporating AI in education, with 81% of the teachers in the US agreeing that AI education is important but countries like Africa still struggling to incorporate AI in education because of their infrastructure.
There has also been a rise in AI-related security accidents which can be avoided through proper legislatures and regulations. In the US, federal agencies have put forward 59 AI-related regulations in 2024, while on global level AI legislations have increased 21% in 75 countries.
Read next: They Look Real, Sound Smart, and Don’t Exist — AI Experts Are Fooling the News
According to the report, developers have made significant progress in AI technology over the past year and have gotten record-breaking scores on benchmarks like GPQA, MMLU, and SWE-bench which were developed in 2023 to test the limits of AI. Many advancements in AI-generated videos and AI agents have also made many tasks a lot easier. The US has the top-performing AI models but China is catching up as well. In 2024, 40 AI models were published by different organizations in the US as compared to 15 from China and 3 from Europe. Chinese models are showing a lot of progress and are competing with the US models on benchmarks like HumanEval and MMLU.
China is quickly gaining ground in AI because of the efficiency, affordability, and accessibility of its AI systems. A report found that the cost of inference for a system that works like OpenAI’s GPT-3.5 has dropped 280 times over the past two years, and the hardware costs have decreased 30% in the last year. 40% less energy is being used by AI models as compared to a year ago and more powerful models are now being used. Many regions like Southeast Asia, the Middle East, and Latin America are also confident in launching their own powerful LLMs. However, the report also found that even the most powerful AI models are struggling to reach human reasoning skills. AI can do well on tasks like the International Math Olympiad but still faces difficulty when it comes to performing tasks which require logical reasoning.
AI is also continuing to attract billions of dollars of investment, with the US having the largest share. American investors invested $109 billion for AI development, as compared to China’s $9.3 billion and UK’s $4.5 billion in 2024. Generative AI has gotten $33.9 billion in private investment, which is an 18% increase from 2023. The report found that 78% of the global enterprises have deployed AI in their workplaces in 2024, which is a 55% increase from 2023.
AI has also made a big impact on society where it is now being incorporated in transportation and healthcare, with FDA approving 950 AI-powered medical devices. AI vehicles like Waymo’s self-driving taxis are also getting about 150,000 rides per week. However, the public opinion in the US about AI is still cautious, with only 39% of Americans thinking of AI as beneficial as compared to more than 75% of people from Thailand, China, and Indonesia who think of AI as advantageous. On the other hand, there is still a big gap when it comes to incorporating AI in education, with 81% of the teachers in the US agreeing that AI education is important but countries like Africa still struggling to incorporate AI in education because of their infrastructure.
There has also been a rise in AI-related security accidents which can be avoided through proper legislatures and regulations. In the US, federal agencies have put forward 59 AI-related regulations in 2024, while on global level AI legislations have increased 21% in 75 countries.
Read next: They Look Real, Sound Smart, and Don’t Exist — AI Experts Are Fooling the News
