DiscussionOpinion

AI Series: Live From New York

Moody's Talks - Inside Economics47m 12s

A panel of economists from Moody's Analytics discusses four AI scenarios for the economy: a baseline graceful integration, a stock market correction, job dislocation, or an upside boom. While they largely agree AI won't cause job apocalypse, they assign significant probabilities to market corrections and express concern about uniform thinking across the industry.

Summary

Mark Saini, Chief Economist of Moody's Analytics, hosts a live podcast with co-hosts Marissa and Chris to synthesize insights from six previous AI-focused interviews. They discuss four possible economic scenarios resulting from AI adoption.

The baseline scenario assumes AI integrates successfully with stock prices remaining flat while earnings catch up, 2% economic growth, fading inflation, and modest productivity gains of approximately 0.5% annually in total factor productivity. Benefits would gradually distribute beyond tech companies to broader sectors.

Marissa expresses increased pessimism about near-term prospects despite long-term optimism. She cites massive capital commitments ($500 billion over 10 years from companies like Anthropic to major cloud providers), uncertain revenue generation, recent stock market declines, and survey data showing only 12% of CIOs report material business transformation from AI spending. She assigns 40% probability to baseline, 35% to stock market crash, and 15% to job dystopia.

Chris presents a more moderate view, expecting some industry shakeout similar to past tech cycles (railroads, automobiles, computers) with consolidation but no severe recession. He assigns 50% to baseline, 30% to stock market correction, 5% to job dystopia, and 15% to upside scenario. He emphasizes demographic tailwinds from boomer spending and questions the strength of wealth effects on consumption.

Mark argues the baseline forecast lacks recession despite panelists' pessimism, explaining that reliable recession timing and severity are difficult to model. The panel discusses potential catalysts for corrections: earnings misses from hyperscalers, disruptions in private credit financing of AI infrastructure, cyber events, regulatory intervention prompted by security incidents, and intellectual property litigation (like the New York Times suit against OpenAI).

Regarding job losses, the panel largely dismisses the apocalyptic scenario. They note that general-purpose technologies historically create new jobs around compliance, security, and management. Discussions with prior guests (David Autor from MIT, Daryl Spence from Capital Group) supported this view. However, they note potential structural shifts: companies may hire more senior positions to manage AI systems rather than entry-level roles, and employment data shows weakness in specific sectors (title insurance, claims adjustment) traceable to non-AI factors.

Key data points discussed include Eric Zarian's RAMP spending data showing 60% AI adoption among his platform's customers, with the top 1% spending $7,000 monthly per employee versus median spend of $600. Anthropic's S1 filing revealed two customers represent 20% of revenue, indicating highly concentrated adoption among power users. The Gartner survey of 2,000 CIOs showed 88% unable to confirm material AI returns on investment.

The panel expresses concern about groupthink consensus that AI will succeed without major disruption, noting that uniform opinion across economists and industry participants suggests systematic misjudgment is likely. Marissa and Chris shift perspectives during the conversation, with Marissa becoming more pessimistic and Mark becoming more optimistic, reflecting genuine uncertainty about outcomes.

About this episode

Inside Economics closes out its AI series with a rare live taping at the Economic Outlook Conference in New York City. Drawing on lessons from earlier episodes—and plenty of input from the audience—Marisa, Cris, and Mark weigh AI's promise against its perils. The verdict: fears of mass job losses are fading, replaced by worries that AI is inflating a bubble in stock and credit markets.

Key Insights

  • Marissa reports that in a Gartner survey of 2,000 chief investment officers, only 12% stated AI has materially transformed their business, while 88% reported no confirmed benefits or inability to measure returns despite billions in AI spending.
  • Eric Zarian's spending data shows a K-shaped distribution where the top 1% of customers account for 80% of AI spending, with the highest spenders allocating $7,000 monthly per employee compared to a median of $600, indicating AI adoption remains concentrated among power users.
  • Anthropic's preliminary IPO filing revealed that two customers represent 20% of total revenue while committing the company to $500 billion in spending over 10 years to major cloud providers regardless of revenue outcomes, creating financial inflexibility.
  • Chris argues that demographic shifts in boomer spending behavior—spending wealth with grandchildren rather than bequeathing it—provide a consumption floor that would cushion the wealth effect from a stock market correction, making downside recession scenarios less likely.
  • The panel identifies concentrated financing risks in AI infrastructure where companies engage in circular financing arrangements (trading chips for data center investments and loans), creating ecosystem vulnerability if any major player faces financial distress.
  • Mark notes that reliable economic forecasting cannot incorporate recession timing and severity without specific data requirements, creating a structural challenge in modeling AI correction scenarios despite panelists' expressed pessimism.
  • Torsten Slok's framing emphasized that while the economy has strong momentum over the next 6-12 months, the longer-term outlook became shakier upon detailed examination, creating ambiguity about how long baseline positive conditions persist.
  • Marissa observes that increased reading about AI capital commitments and leverage has shifted her from month-ago optimism to current concern, demonstrating how deepening due diligence on financial arrangements is driving pessimistic scenario reassessment.

Topics

AI economic scenarios and probability distributionsStock market valuation risks in AI sectorLabor market impacts and job displacementCapital expenditure commitments and financing risksProductivity growth potential and measurementCorporate AI adoption rates and ROIRegulatory and security risksDemographic effects on consumption and wealth

Transcript

Welcome to Inside Economics. I'm Mark Saini, the Chief Economist of Moody's Analytics. I'm joined by my two trusty co-hosts, Marissa Dina Telling and Chris Therese. Hi, guys. Hi, Mark. How are you doing this afternoon? Hey, this is a special podcast, right? I mean, because we have a live audience. We're taping this at Moody's HQ, Seven World Trade Center. We had a great day. Right, guys? This is the audience. And this is the end of the day. We're in the middle of happy hour. Today, before we dive into the topic at hand, and that's artificial intelligence, because we ran a series of podcasts around AI, five, six different guests, and talked about different aspects of AI.…

Full transcript available for MurmurCast members

Sign Up to Access

More from Moody's Talks - Inside Economics

Get AI summaries like this delivered to your inbox daily

Get AI summaries delivered to your inbox

MurmurCast summarizes your YouTube channels, podcasts, and newsletters into one daily email digest.