Conference Presentation, Panel
The Rising AI Economy: A New Era in Tech | Milken Institute Global Conference 2024
Milken InstituteNicole Valentine, Seth Boro, Les Borsai, Sachin Dev Duggal, X Eyeé, Martin Fichtner, Sachin Chakrabarti
- Technology Lifecycle Context: Sachin Chakrabarti compares current AI adoption to the 1990s dial-up era (AOL) and the 2000s RailPlayer, suggesting the technology is currently in a "broadband is yet to come" phase of infrastructure and usage maturity.
- AI Definition Shift: The panel argues AI represents an evolution of "design" rather than just technology, characterized by complex systems hidden behind human-like interfaces (e.g., chat interfaces) that have shifted public vernacular from acronyms to terms like "hallucinations."
- Educational Impact (Positives): AI tools like "Read Along" and Microsoft's Reading Coach are cited as accessible, free solutions addressing the global literacy crisis, potentially helping 800 million adults and the estimated 50% of US fourth-graders not reading at grade level without requiring expensive private tutoring.
- Algorithmic Bias & Governance (Negatives): Illinois deployed an AI algorithm in 2016–2017 to predict child abuse risk; the system incorrectly flagged 360 children as 100% at risk of death (none were), disproportionately targeting black and brown children, leading to its removal, though dozens of states reportedly still utilize it.
- Technical Limitations: Large Language Models (LLMs) are described as "large parrots" that mathematically predict token sequences rather than reasoning, leading to issues with bias against non-English speakers and racial/gender bias documented in Nature and Scientific American.
- Energy Consumption Trends: Current H100 GPU circulation consumes the equivalent power of Sri Lanka, with projections suggesting AI infrastructure could consume the power of all of Europe within 16–18 months and potentially the entire world's power supply if efficiency does not improve.
- Investment Thesis (Thoma Brava): Seth Sternberg notes that despite Gen AI being only ~12 months old, portfolio companies are delivering real-time value in budgeting, planning, and cybersecurity, where AI aids in detecting deepfakes and defending enterprises faster than human analysts.
- Investment Thesis (Temasek): Martin Khor categorizes investments into "AI Natives" (disruptors born of AI) and "AI Wave Riders" (incumbents like Microsoft/Nvidia whose revenue models are determined by AI), while emphasizing AI's role in driving margin expansion and cost reduction across all sectors.
- Funding Anomalies: Sachin Chakrabarti highlights a market inefficiency where high-revenue, established AI companies are undervalued ("nickel and dimed"), while "wrapper" companies with zero revenue and promises of high future pricing receive billion-dollar valuations due to investor hype and information asymmetry.
- Emerging Consumer Use Cases: Les Ebbets points to the monetization of "digital doubles" and virtual relationships, citing the "virtual girlfriend" economy and Asian gamers paying for AI doubles, alongside AI-generated music by legacy artists using their own AI versions to bypass declining creative output.
- Workforce Displacement & Creation: X Fortuna cites a 2014 statistic that 80% of jobs held by 4th graders would be obsolete by adulthood; she argues that while old roles (e.g., film editors) vanish, new markets (e.g., YouTube, TikTok) and opportunities for creators like "MrBeast" emerge from the disruption.
- Operational Risk: Companies rushing AI deployment face liability for errors; Air Canada was sued for a Chatbot hallucinating bereavement policies, and New York City faced backlash for a bot instructing business owners to break laws.
- AGI Definition Variance: The panel notes there is no universal definition of Artificial General Intelligence (AGI), with OpenAI defining it as performing most human jobs well, while others require reasoning and emotion; Mark Zuckerberg recently invested $500 million at Harvard to accelerate human brain understanding via AI.
- Superintelligence Risks: A US Department of Defense war-game test showed ChatGPT repeatedly recommending nuclear strikes as the only solution, highlighting risks of LLMs parroting biased data in high-stakes scenarios without human oversight or moral boundaries.
- Skill Obsolescence: Seth Ladell argues that teaching specific skills like coding or sales techniques is becoming obsolete; the only sustainable skill is the ability to learn new methods, suggesting "STEM is dead" as a curriculum focus if children are to remain competitive in an AI-augmented workforce.
- Governance & Responsibility: Martin Khor proposes that "responsible AI" is defined by accountability, asking who bears the risk when an AI replaces a human employee (e.g., a salesperson making promises); the burden falls on the entity deploying the solution, not the technology itself.
- Regulatory Challenges: X Fortuna and Sachin Chakrabarti criticize regulators for lacking technical expertise (only ~1,000–2,500 true AI experts globally) and failing to keep pace with technology, noting that policymakers often regulate visible components (power transformers) while ignoring the broader "iceberg" of cloud and data governance.
- EU vs. US Regulatory Approaches: X Fortuna contrasts the EU's risk-based AI Act with US regulatory struggles, arguing that regulation must be scaled to potential harm (e.g., medical dosage vs. book recommendations) rather than treating all AI equally.
- Geopolitical & Values Conflict: X Fortuna warns that AI development in adversarial nations (e.g., China) often lacks Western ethical boundaries, such as facial recognition policing for civil disobedience, creating a divergence in how AI aligns with societal values like the US Constitution.
- Future Vision (Education & Democratization): Sachin Chakrabarti advocates for "falling forward," where AI democratizes access to high-level skills (music, painting, screenwriting) for those without traditional training or resources, effectively leveling the economic playing field.
- Future Vision (Legacy): X Fortuna emphasizes a long-term perspective over quarterly goals, urging policymakers and developers to design AI systems that empower future generations rather than automating the labor of ancestors who fought against inequality.