Conference Presentation, Panel
On AI: Discourses on Application, Convergence, and the Future of Humanity | Asia Summit 2024
Milken InstituteElla Tan, Gary Liu, Lila Ibrahim, Phoram Mehta, John Quinn, Rob Schimek, Shan Shan Wong
- Generative AI is projected to grow at over 40 percent over the next 10 years, reaching a market value near or exceeding $1 trillion by 2032, while driving an incremental $280 billion in new software revenues as enterprises adopt applications to boost productivity.
- Training costs for large language models are expected to continue rising exponentially, having increased from $1,000 in 2017 to nearly $80 million for ChatGPT 4 and approaching $200 million for Google's Gemini Ultra, which may result in a market where only a few companies provide the foundational models while hundreds of thousands to millions of derived applications enter daily use.
- Regulatory environments remain uncertain, with a comprehensive AI bill currently pending in California that may apply next year, though full global materialization of regulations across jurisdictions is expected to take years, potentially establishing location-specific rules as worldwide defaults.
- In the insurance sector, AI applications are expected to help close the current $7 trillion protection gap and narrow disparities between haves and have-nots caused by climate change and geopolitical risks, while also enabling businesses to better predict and monitor fraud, with smaller firms anticipated to rely on larger entities for ethical frameworks in the near term.
- Generative media is expected to transform visual arts by lowering barriers to creating three-dimensional renders from two-dimensional images and reducing the need for expensive studio time, while AI is also predicted to act as a co-pilot to enhance accountability and transparency in decision-making processes.
- Legal and medical applications are expected to focus on assistance rather than replacement, with AI aiding in life sciences, cancer research, vaccine development, personalized medicine, and fraud monitoring in Asia, but not replacing human decision-makers in courtrooms or prison sentences despite its predictive capabilities regarding recidivism.
- Over the next decade, AI is expected to facilitate the discovery of 400,000 new materials compared to the previous 40,000, demonstrate the ability to analyze over 200 million proteins, and provide personalized tutoring to every student, thereby fostering curiosity and problem-solving to make the world "more human."
- Global collaboration is expected to continue through open-source initiatives and big tech companies to accelerate AI development, with hopes that AI will eventually assist in electing leaders based on enhanced emotional or intellectual intelligence and support sectors such as motor security, climate science, and health.