AI’s next phase will depend less on smarter models and more on whether companies, governments and communities can build the infrastructure, workflows and legal frameworks to support them. That was the central message from Norm AI’s annual Central Park AI Forum, according to a summary published by the company.
Norm AI hosted the event on September 24 at The Plaza in New York. More than 200 leaders from business, technology, law and government attended, and the focus moved from speculation to the operational decisions organisations already face. Norm AI noted that in the twelve months since its previous forum, AI agents have become embedded in business workflows, compute demand has surged, and lawmakers have begun weighing the implications of an agentic economy.
The first theme was individual autonomy. Norm AI founder and CEO John Nay argued that the American legal tradition, which balances personal freedom against limits on harmful conduct, offers a guide for AI progress. Khosla Ventures founder and CEO Vinod Khosla described AI as a potential equaliser that could make medical, financial, educational and legal expertise widely available at low cost. He also called for policies that support people through economic disruption and into more fulfilling work.
Physical limits emerged as a pressing concern. Mike Forman pointed to record data centre demand, with agentic development expected to push compute requirements higher still. Power connections, transmission capacity, land, permits and skilled labour can each delay projects, and local communities are increasingly scrutinising water, electricity and employment impacts. New York State Assembly member Micah Lasher said government will need to broker agreements between developers and communities.
For enterprises, Norm AI’s summary suggests the bottleneck is organisational rather than technical. Firms must map processes, define escalation points and give agents secure access to fragmented data. Salesforce president and chief legal officer Sabastian Niles outlined efforts to build an AI-native legal, compliance and risk function. Morgan Stanley chief legal officer and chief administrative officer Eric Grossman discussed how AI can standardise legal processes and preserve institutional knowledge.
Judgment was another recurring theme. Coatue founder Philippe Laffont distinguished judgment from expertise in a conversation with Anthropic chief economist Peter McCrory, who cited data showing that domain experts guiding Claude Code perform better, particularly on complex tasks. Laffont warned that automating entry-level work could erode the pipeline of future senior decision-makers. Stephen Wolfram added that sophisticated systems behave unpredictably, so human judgment remains essential.
Finally, Norm AI framed law as infrastructure for an agentic economy. Delaware Secretary of State Charuni Patibanda-Sanchez and former Thomson Reuters CEO Tom Glocer discussed Delaware’s proposed sandbox for Agent-Managed Companies, in which AI agents could run human-owned entities under registration, disclosure and oversight rules. Norm AI is helping the Delaware Department of State build the sandbox’s infrastructure.
Norm AI concluded that America must expand compute, redesign workflows, protect the routes through which people develop judgment, and create legal structures fit for AI agents.
For more insights from the panel, read the full story here.
Read the daily FinTech news
Copyright © 2026 FinTech Global
Copyright © 2026 RegTech Analyst
