Artificial intelligence is evolving at a breathtaking pace. A few years ago, OpenAI’s work such as his ChatGPT and Google Bard were limited to outrageous discussions at technical roundtables. But a lot has changed in the last few months. AI technology is developing at breakneck speed, leaving some individuals and organizations speechless with awe and anxiety.
Stanford University recently released its 2023 AI Index report. The AI Index is a Stanford University Human-Centered Artificial Intelligence (HAI) initiative led by the AI Index Steering Committee. Reports essentially track, collate, extract, and visualize data related to AI. According to the agency, the report will enable “decision makers to take meaningful action to advance her AI responsibly and ethically, with humans in mind.”
The proliferation of AI-powered technologies is clear, but the report made some surprising discoveries about the rapid adoption of the technology. Interestingly, the report notes that his 1,800 key figures, including billionaire Elon Musk, scientist Gary Marcus, and Apple co-founder Steve Wozniak, have all worked together to develop a system more powerful than GPT-4. announced when it signed an open letter calling for it to be suspended for six months. .
Here are the key takeaways from Stanford University’s 386-page 2023 AI Index report.
bigger and faster
AI models are getting smarter, bigger, and faster. AI models are now training on a million times more data than he did 10 years ago. According to the report, until 2014, the most important machine learning models were released by academia, but now appear to have been superseded by industry. In 2022, the 32 significant machine learning models created by industry contrasted with just his three machine learning models created by academia. The widening gap is due to the large amounts of data, computing power, and funding required to build AI systems. These are things that the industry is well endowed with compared to non-profit organizations and academia.
AI models are much more capable today than they were two years ago. This is due to technical performance benchmarks across language or text, image and video evaluations. The report states that AI consistently reports state-of-the-art results, although many benchmarks show modest year-over-year improvements. Also highlighted was the release of his suite of new, more comprehensive benchmarks such as Big-bench and HELM.
AI models need more resources
Training large language models is expensive. GPT- (2019) was trained at an estimated $50,000, while Google’s PaLM (2022) cost him $8 million. Interestingly, for the first time in a decade, private investment in AI has declined year-over-year. According to the report, global AI private investment in 2022 will be $91.9 billion, down 26.7% compared to 2021. The total number of AI-related fundraising events and the number of newly funded AI companies are down. According to the report, private investment in AI in 2022 will be 18 times higher than in 2013.
More companies advocate for AI
The percentage of companies implementing AI has more than doubled since 2017. And companies that have integrated AI report reduced costs and increased revenue. “Organizations that have adopted AI report significant cost savings and revenue increases.”
Growing demand for AI jobs and job satisfaction among co-pilots
Demand for AI-related professional skills has increased significantly across America’s industrial sectors, according to the report. AI-related job postings increased from 1.7% in 2021 to 1.9% in 2022. US employers are increasingly looking forward to hiring people with AI-related skills, according to the report. More AI PhDs choose jobs in the private sector (65%) compared to academia (28%) or government (1%).
Additionally, AI co-pilot tools are making people more productive. Of his Copilot users on Github, 88% of him reported being more productive, and 74% said he was able to focus more on satisfying work.
Increase in AI abuse
Even as companies scramble to develop new AI technologies, there are countless examples of misuse of these technologies. The report points to an increase in AI-powered deep fakes and security breaches. Overall, the number of AI incidents and disputes has increased 26-fold since 2012. These incidents stem from the growing use of AI technology and growing awareness of the potential misuse of AI technology. The report also suggests that AI systems can have a serious impact on the environment.
Policy makers enthusiastic about AI
The report analyzed the legislative records of 127 countries and showed a staggering increase in the number of AI-related bills passed, from 1 in 2016 to 37 in 2022. Procedures increased 6.5 times compared to 2016.