Conference Presentation, Panel
Leading in the Generative AI Revolution | Milken Institute Global Conference 2024
Milken InstituteSara Eisen, Greg Brown, Arvind Krishna, Thomas Kurian, Janet Truncale, Sharmini Peries, Janet Napolitanovich, Thomas Kaplan, Nicole Sullivan, Arvind K.
- The AI landscape is characterized as being in its "first innings," with cloud technology potentially in its "fifth or sixth innings" and other emerging technologies still in early development phases.
- Investment in generative AI is expected to remain a primary focus for major organizations like EY and Google, with EY targeting its "almost 400,000" employees for training and Google having built state-of-the-art models over the "last three years."
- Strategic plans include shifting from "productivity gains" and "use cases" to "growth opportunities" and "pilots" in high-opportunity, low-risk sectors, with EY anticipating a pivot to live deployment in most areas except finance, which may move "a little slower" due to regulatory concerns.
- Specific use cases involve Motorola Solutions using AI to detect security threats like "stolen vehicles" or "brandished weapons," aiming to reduce 911 response times by one minute to potentially "save 10,000 lives" and process evidence "10 to 15 times faster."
- Technological capabilities are expected to expand to process "over an hour of video at one shot" or "70,000 pages of information," with Google introducing "mathematically constrained" models for tasks like credit scoring and "grounding" features to cite sources from the "entire internet" or proprietary data.
- Google plans to release verification technology to distinguish between human and AI-generated content, while IBM aims to provide "reasonably sized" models that can be deployed internally at lower costs to prevent "IP leakage."
- Demographic shifts in "Japan, China, UK, Italy, Spain, France, Germany," and the "bulk of the developed world," where birth rates are "below substitution rates," are expected to drive the replacement of human workers with "digital workers" for "clerical" and "matching" tasks, though "human interaction" remains irreplaceable.
- Efficiency gains from AI coding are projected to allow teams to write code "20% faster," resulting in more product differentiation rather than job cuts, with semiconductor optimization expected to yield a "100, maybe 1,000 times advantage" in two years.
- Future applications include AI tutors with a "small brain" to assist children, real-time crime centers for locating missing persons, and edge-based systems to guide individuals to safety during active shooter situations involving a "brandished rifle."
- Workforce expectations indicate that while AI may augment humans and reduce paperwork friction by "3%, 4%, 5%," it cannot yet replicate "EQ," which leaders expect robots to only demonstrate at a "first or second grade level."
- Regulatory outlooks anticipate a "lighter touch" initially with "principles" for data usage, with leaders warning that "overregulation" could inhibit innovation, while copyright and "fair use" disputes are expected to be resolved through "legislation" or "courts."
- EY expects to eventually open its proprietary model, currently processing "up to 25 million prompts," to partners, while maintaining strict data protection to ensure "trust" in sectors like healthcare where AI could predict illness.
- The financial sector is currently utilizing AI primarily in "FP&A" and "forecasting," with public safety and healthcare expected to lead in live deployment pending the establishment of trust and resolution of privacy concerns.
- Geopolitically, the US is expected to maintain a lead in AI assets, having spent "nine times more on AI investment than China last year," with Google seeking to ensure models are "culturally appropriate" for global markets.
- Key risks identified include "misinformation" as the "biggest danger," potential bias due to uneven data availability, and the need for "fulsome" data to build system trust, alongside the uncertainty of whether AI adoption will be a "when" rather than an "if" question regarding scaling.
- Economic projections suggest that the IMF forecast of "40% of the jobs" being eliminated may shift toward increased "productivity," "growth," and "GDP," with EY aiming to generate the "next 50 billion" dollars using significantly less growth in traditional talent metrics.