Panel, Fireside Chat, Conference Presentation
AI and the Future of Digital Transformation | Future of Finance 2026
Milken InstituteNicole Valentine, Jo Jagadish, Karen Kornbluh, Eric Mandl, Igor Tulchinsky, Kip Wainscott
- Panel Context & Objective: The session focused on steering the next industrial revolution in AI through diverse perspectives (finance, policy, tech, academia) to balance benefits against harm, specifically addressing economic opportunities, workforce disruption, industry applications, and national security.
- TD Bank Strategy (Joe Jagadish):
- Prioritizes a "human-in-the-loop" evolution toward "humans as supervisors" to ensure human judgment remains central despite AI efficiency gains.
- Achieved an 80% adoption rate for a new internal AI knowledge management tool in contact centers within the first few weeks of deployment.
- The tool has reduced average handle times and freed up banker capacity for higher-quality customer engagement and fraud management in the payments ecosystem.
- Plans to expand testing from internal use cases to customer-facing applications over the coming years.
- Policy & Economic Outlook (Karen Kornblum):
- Views current AI developments as a continuation of the digital revolution, noting that productivity gains historically lagged the initial technology rollout by a decade before accelerating.
- Warns that professional services (law, accounting, architecture) face existential disruption similar to taxis vs. Uber, requiring a redefinition of "moats" and data assets.
- Highlights the need to unlock private data for pharmaceutical breakthroughs, citing the current barrier where failed drug trial data remains private.
- Suggests a potential political pivot where AI regulation becomes a central 2028 election issue, driven by public demand for guardrails against mass surveillance and lack of human oversight.
- Notes that while the industry lobbies for zero regulation, polling indicates the public desires a "pendulum swing" toward establishing specific guardrails and reducing industry influence.
- Market Dynamics & Investment (Eric Mandel):
- Projects AI economic spend to reach between $4–5 trillion over the next decade, potentially rising to $10–15 trillion or the size of current US GDP ($29T), far exceeding the current ~$700B hyperscaler spend.
- Observes that 90% of all data on earth was created in the last 24 months, with the doubling cycle shortening from four years to two due to AI generation.
- Argues that despite market anxiety, historical tech shifts invariably expand economic output and workforce size, creating more jobs than they displace.
- Emphasizes that the market currently assumes all technological change happens instantly (extreme discount rates) rather than over a realistic timeline, leading to mispricing of risk and opportunity.
- Confirms that Guggenheim is hiring more engineers and technical staff to manage complex AI-integrated questions, noting that professional services remain the largest component of tech deployment.
- Technology & Venture Scaling (Igor Telchinsky):
- Characterizes current AI progress as a "step function," where processes are now 100x faster, though the organization faces a talent shortage to deploy these capabilities at scale.
- Maintains that quantum computing will not replace GPUs but will solve specific optimization and encryption problems (e.g., cracking RSA codes) 100 million times faster.
- States that WorldQuant's headcount has increased post-ChatGPT, with the workforce shifting toward higher-level tasks while automation handles lower-level work.
- Sets a goal to increase workforce productivity 100x within the year through aggressive automation of discoverable processes.
- Identifies defense as the sector where AI is most relevant, arguing that minimal code-level guardrails are necessary to maintain global competitiveness against nations with fewer regulations.
- Enterprise Policy & Trust (Kip Wayman, JPMorgan Chase):
- JPMorgan operates as a "procurement node" with $20B annual tech spend and 300,000+ employees, allowing it to evaluate policy from three vantage points: deployer, purchaser, and investor.
- Achieved $1B in fraud loss prevention in the past three years, projecting an additional $3.7–3.8B in savings over the next four years via AI models.
- Deployed LLMs to 250,000 employees, with 160,000 weekly active users reporting an average of 4 extra productivity hours per week.
- Launched an "AI threat modeling co-pilot" to identify cyber risks earlier in the software development lifecycle.
- Highlights a critical trust deficit: consumer trust in AI is declining year-over-year, which threatens enterprise adoption and necessitates robust, technology-neutral policy frameworks.
- Warns that "agentic commerce" (AI agents making transactions) currently lacks standard-setting bodies, creating friction for banks as the primary point of dispute resolution.
- Workforce Displacement & Creation:
- Cites data showing a net increase in "threatened" roles since November 2022: +3M white-collar roles, +7% engineers, +21% paralegals, and +10% radiologists in the US.
- Draws parallels to the ATM era, where branch banking persisted and evolved rather than disappearing, suggesting new job categories will emerge to support the AI economy.
- Notes that the "law of unforeseen consequences" often leads to job creation in tech infrastructure and support services rather than net unemployment.
- National Security & Governance:
- Highlights ongoing tensions between the DoD and tech firms (e.g., Anthropic) regarding mass surveillance of US citizens and the removal of human-in-the-loop requirements.
- Warns that countries without significant data privacy regulations (e.g., certain small nations or China) may rapidly outpace regulated nations in defense capabilities if they can access global data freely.
- Recommends that guardrails belong in legislation rather than embedded in code to avoid creating "bad AI" that is uncompetitive.
- Forward-Looking Actions & Resources:
- Education: Speakers urge active "hands-on" learning (building apps, using AI for personal tasks like bill auditing) rather than passive consumption.
- Policy Monitoring: Advises following geopolitical trends (sovereignty, export controls) and specific resources like the "China Talk" podcast for insights on international AI competition.
- Skill Development: Recommends using tools like NotebookLM and developing personal thesis statements on AI ethics and regulation, as the "right answer" to governance is currently unknown.
- Standardization: Calls for the immediate creation of new standard-setting bodies to govern agentic commerce and cross-platform transaction disputes.