AI's Next Chapters: What to Expect from A Transformative Technology

Robert Kirkpatrick (UN Global Pulse)

• Identifies deepfake technology (specifically GANs used to create plausible video content) as a primary long-term risk that could undermine the rule of law by creating unprovable "fake news" evidence, potentially inciting war or framing innocent individuals.
• Highlights the dual nature of AI: while $500 drones can transform subsistence farmers in Africa into entrepreneurs, they also pose a risk of displacing labor if not managed correctly.
• Proposes a "socioeconomic equivalent of meteorology" using real-time data to predict and prevent crises, citing an 89% accuracy rate in predicting food insecurity in Rwanda based on mobile airtime spending patterns.
• Details specific predictive applications for disaster response, including using ship movement signatures in the Mediterranean to detect refugee rescue events and analyzing mobile network data flows to model population displacement and disease spread (cholera, dengue).
• Emphasizes the responsibility to use "Data Philanthropy" to anonymize data from mobile and financial sectors to monitor global well-being without compromising individual privacy.
• Advocates for balancing privacy risks against the "opportunity cost" of data non-use, noting that preventable deaths in healthcare and lack of disaster response are current harms caused by data restrictions.
• Supports the development of "differential privacy" technology to allow data sharing for purposes like anti-money laundering without compromising underlying individual data, viewing it as a solution that could complement regulations like GDPR.

Mike (DayCred Labs)

• Defines the company's mission to build a digital model of the planet using AI and big data to better understand complex planetary phenomena, specifically global agriculture and energy production.
• States the goal is to tackle global challenges like climate change and world hunger by moving beyond simple data collection to modeling complex financial resources.

Velarde James (WorldClock Predictives)

• Positions his company to apply quantitative finance signal detection methodologies to alternative data for social good, such as predicting which patients will benefit most from clinical trials.
• Highlights a specific use case of analyzing social behavioral data to predict PTSD or suicide risk among veterans returning from Iraq.
• Identifies himself as a "techno-optimist" who believes the greater risk to the developing world is the underuse of AI rather than misuse, citing AI's potential as a "first multiplier" for nations lacking technological independence.
• Predicts significant disruption in healthcare within the next five to six years, specifically in disease prediction, genomics, and understanding lifestyle-related health risks.

Vijay Dorada (Spark Cognition)

• Focuses on industrial AI applications in aviation, oil and gas, and power generation, with a joint venture with Boeing to build operating systems for autonomous flying vehicles.
• Describes a specific use case involving the monitoring of 70 critical assets on offshore oil drilling platforms to detect "unknown unknowns" by correlating data from thousands of sensors to prevent failures.
• Proposes a "pseudo-autonomous" power generation model where facilities in developing nations can be managed remotely from the West to solve workforce training shortages and accelerate economic momentum.
• Views AI as a force multiplier for safety and efficiency in industrial settings, aiming to prevent accidents and optimize throughput in sectors where skilled labor is scarce.

Erkin Abilov (Behavox)

• Uses machine learning to mine employee behavioral signals from communications data (emails, chats, calls) to predict business outcomes like compliance issues, performance, and HR motivation, primarily within financial services.
• Cites a client (Citigroup) generating 20 million emails daily as an example of the volume of sensitive data requiring automated analysis.
• Addresses the "chicken and egg" problem of training data privacy by developing pre-trained models in-house to deliver ready-made software to clients, ensuring sensitive data never leaves the secure environment.
• Notes that GDPR and privacy regulations actually accelerate the adoption of their technology, as clients require automated solutions to meet "right to be deleted" and data governance mandates.
• Describes the shift from "humans vs. AI" to "cyborgs," arguing that enabling employees with AI tools increases individual productivity by 50 to 100 times, necessitating significant workforce upskilling.

Panel Consensus on Data and Policy

• Bias Mitigation: Strategies include "idea arbitrage" by deploying global teams (e.g., Hungary, Vietnam) to build diverse predictive models that are merged into "metamodels," and using indigenous speech recognition to capture data from populations excluded from digital networks (e.g., radio calls in Uganda).
• Regulation: The panel agrees that while media often overhypes General AI, current policies must be grounded in understanding Narrow AI to avoid decisions based on ignorance; however, some (like Abilov) argue that compliance regulations drive necessary investment in AI solutions.
• Global Competition: Noted that China invested $20 billion in AI in 2017, with a projection to reach $70 billion by 2025, creating a strategic imperative for the US to define its own long-term strategy.
• Demographics: Cited 2016 data showing India (78M) and China (77M) surpassed the US (67.5M) in total graduates, with China producing 4.67M STEM graduates compared to the US's 568,000.
• Job Displacement vs. Augmentation: Consensus that while AI will displace specific roles (particularly in white-collar legal and financial analysis), the primary outcome should be the amplification of human intent rather than displacement of agency, requiring a shift in educational focus toward computer science.

Future Outlook and Industry Impact

• Healthcare: Universally identified as the primary sector for immediate, life-altering disruption through disease prediction, genomics, and early intervention.
• Financial Services: Predicted to face the most rapid transformation (5-6 years) in market-making, compliance, and risk analysis, particularly in OTC instruments and fixed income.
• Climate & Agriculture: Anticipated impact through more efficient global supply chains, monitoring of externalities like pollution, and AI-driven optimization of agricultural resources.
• Main Street Integration: The panel expects AI to become an invisible infrastructure (like Alexa) in daily life, with the most tangible shifts occurring in productivity tools for business services rather than overt consumer-facing "robots."