Panel, Conference Presentation
AI's Next Chapters: What to Expect from A Transformative Technology
Milken InstituteBecky Peterson, Erkin Adylov, Vijay Doradla, James Golden, Mark Johnson, Robert Kirkpatrick
- Robert Kirkpatrick warns that unchecked convergence of AI with other technologies, specifically deepfakes, could endanger society, potentially inciting wars and undermining the evidentiary foundations of the rule of law by creating unprovable fake news.
- Predictions include AI applications, such as $500 drones, transforming subsistence farmers into entrepreneurs within a timeframe not specified, while society is expected to face "incredible opportunity costs" and preventable deaths in health, education, and disaster response if data usage is not accelerated.
- The United Nations is expected to create a "socioeconomic equivalent of meteorology" within the coming decade, utilizing real-time behavior data and predictive algorithms to scale prevention programs, recognize discrimination in "a thousand languages," and generate real-time maps of conflict and service denial.
- Specific predictive capabilities are forecasted to achieve 89% accuracy in predicting food spending from mobile data in Rwanda, while predictive analytics will enable governments to forecast population flows in disaster zones, anticipating specific numbers like "2,000 people there and 800,000 more" upon arrival.
- Differential privacy technology is predicted to resolve data sharing challenges within the "next 5 years or 10 years," and while GDPR may not hinder progress, it is expected to help privacy protection if paired with such new technologies.
- Vijay Dorada anticipates an exponential increase in productivity and problem-solving for companies adopting AI, warning that countries not using AI within "year two" will fall significantly behind due to tenfold annual data growth.
- Tech giants like Google, Facebook, Apple, and Amazon are expected to become "stronger and stronger" by leveraging unique data assets, while the "chicken and egg" problem of training data access will be solved by companies shipping "ready-made models" from on-premises to clients.
- The US government is predicted to require a long-term AI strategy over the "next couple of decades," contrasting with China's AI spending which rose from $20 billion in 2017 to an expected $70 billion by 2025.
- Regulations are expected to accelerate AI adoption and stabilize financial systems by bridging black box machine learning gaps in policing, as human policing of financial institutions is described as "humanly impossible."
- Companies like Facebook and Twitter are expected to face regulation and accountability for promoting terrorist content, while a future scenario where "robots and all of AI are owned by 2 trillionaires" is identified as a negative outcome to be avoided.
- Automation is projected to drive economic prosperity despite the risk of joblessness in post-conflict environments if deployed without strategy, and AI is expected to scan 100,000 legal pages to find loopholes within the "next couple of years," putting legal jobs at risk.
- Healthcare, specifically the prediction and prevention of disease, is expected to be the area where "most people will experience the disruptions from AI" in 20 years, leveraging genomics and epigenetics for longer, healthier lives.
- James Velarde predicts AI will act as a "first multiplier" for the developing world to become competitive, with applications ranging from optimizing traffic lights via pollution modeling to predicting PTSD or suicide risks in veterans.
- Wall Street is expected to see "complete disruption" of individual brokers in OTC instruments and fixed income within the "next five years," with efficiency gains projected to drop costs and increase fund availability for the broader economy.
- In industrial sectors, AI is expected to bring autonomy to power plants to improve efficiency, create operating systems for autonomous flying vehicles, and allow power plants to be managed remotely with fewer on-site skilled workers.
- Erkin Abelov predicts healthcare will face disruption first, followed by business services like legal, finance, and accounting in "five to six years," while China and India are expected to produce significantly more STEM graduates than the US.
- The future workforce is expected to require scripting language proficiency, leading to increased demand for up-skilling as AI transforms humans into "cyborgs" who are "50 times, 100 times more productive."