Pope Leo Issues AI Encyclical with Anthropic, BofA Warns of AI Bubble, Trump Scraps AI Oversight Order — AI News Briefing

🗞️ AI News Briefing — May 25, 2026 (06:00 CST)

Top 7 Stories

1. Pope Leo to Issue Landmark AI Encyclical on Human Dignity Alongside Anthropic Co-Founder

In an unprecedented move that bridges centuries of religious tradition with cutting-edge technology, Pope Leo is set to present a major encyclical text on human dignity in the age of artificial intelligence — and he’s doing it alongside a co-founder of Anthropic. According to reports from The Guardian, National Catholic Reporter, and NBC News, the collaboration represents the Vatican’s most substantive engagement with AI development to date and signals a growing recognition that ethical frameworks for AI require input from both moral philosophy and technical expertise.

The encyclical is expected to address fundamental questions about human dignity, agency, and moral responsibility as AI systems become increasingly capable and autonomous. The choice of Anthropic as the Vatican’s partner is particularly notable: the company’s name itself derives from “anthropic principle” — the philosophical concept that the universe’s physical constants seem fine-tuned for human existence. Anthropic’s research focus on AI safety and constitutional AI aligns closely with the Vatican’s emphasis on moral frameworks for technology development. Religion News Service reports that an “unlikely Vatican-Anthropic relationship” has been developing over time, reshaping the broader AI ethics debate.

The collaboration raises significant questions about the role of religious institutions in technology governance. The Catholic Church, with its 1.3 billion adherents and global institutional presence, has historically wielded substantial moral influence on issues ranging from bioethics to economic justice. By partnering directly with a frontier AI lab rather than issuing a standalone critique, Pope Leo is taking a more engaged approach — one that could influence how AI companies think about safety, fairness, and human welfare from the ground up rather than as an afterthought.

2. BofA Warns AI Stock Mania Nearing Historic Extremes — ‘Biggest Bubble Since Railroads’

Bank of America’s chief cross-asset strategist Michael Hartnett has issued a stark warning to investors: the AI-driven stock market rally is approaching what he characterizes as “historic extremes,” comparable to the railroad bubble of the 19th century. According to reports from Seeking Alpha, Moneycontrol, and Investing.com, Hartnett’s analysis suggests that the current enthusiasm around AI stocks has reached levels not seen since the dot-com era, with potentially devastating consequences for latecomers to the rally.

The warning comes as the S&P 500 approaches 7,500, buoyed by AI optimism and hopes for an Iran deal, per MSN’s market reporting. BofA’s fund manager survey reveals a complex picture: while stock market mania appears to be subsiding in some areas, the concentration of capital into AI-related equities remains extraordinarily high. Hartnett’s railroad analogy is particularly pointed — the 1840s railway mania saw massive capital inflows into railway stocks, followed by a devastating crash when reality failed to match the speculative fervor. The implication for today’s AI market is clear: while AI technology itself is transformative, the financial expectations embedded in current stock prices may far exceed what the technology can realistically deliver in the near term.

This assessment stands in stark contrast to another BofA analysis published the same day, which argued that AI’s productivity upside could be 10 times current estimates, according to Investing.com. The contradiction between these two BofA reports captures the central tension in the current AI investment landscape: the technology’s long-term potential is genuinely enormous, but the short-term financial expectations may be unsustainably inflated. For investors, the takeaway is nuanced — AI is real and transformative, but current market prices may already reflect decades of expected progress.

3. Trump Cancels AI Executive Order on Model Oversight After Industry Pushback

In a dramatic reversal, President Trump has canceled the signing of an executive order that would have established federal oversight of AI model development. According to The New York Times, Politico, CNBC, and the Wall Street Journal, the decision came after White House AI czar David Sacks raised concerns from the tech industry about potential overregulation, with Trump stating he “didn’t like certain aspects” of the proposed order.

The canceled order, which had been characterized by some as cybersecurity-focused according to Broadband Breakfast, would have introduced federal oversight mechanisms for AI model development and deployment. The last-minute cancellation — occurring just hours before the scheduled signing — underscores the intense lobbying pressure that big technology companies have brought to bear on AI policy. Politico’s reporting indicates that Sacks played a central role in communicating industry concerns to the President, effectively derailing a regulatory framework that had been in development for months.

This decision has significant implications for the global AI governance landscape. While the European Union’s AI Act has established comprehensive regulatory requirements and China has advanced its own AI governance framework, the United States has now effectively chosen to forgo mandatory federal oversight in favor of industry self-regulation. The cancellation raises questions about whether the U.S. will maintain its leadership role in setting global AI standards and whether the absence of regulatory clarity will ultimately help or hinder American AI companies competing internationally. For AI safety researchers and policy advocates, the canceled order represents a significant setback in efforts to ensure that frontier AI development proceeds with adequate guardrails.

4. Microsoft Reports Expose AI’s Real Cost Problem — Using the Tech Is More Expensive Than Paying Human Employees

A new analysis of Microsoft’s internal reports has revealed a uncomfortable truth about enterprise AI adoption: for many use cases, deploying AI technology is currently more expensive than employing human workers. According to Fortune, Microsoft’s own data shows that the compute costs associated with running AI systems can far exceed the labor costs they’re meant to replace, challenging one of the fundamental assumptions driving the current AI investment boom.

The Fortune report, which echoes earlier comments from a NVIDIA executive who noted that “the cost of compute is far beyond the costs of the employees,” highlights a growing disconnect between AI’s theoretical efficiency gains and its practical economics. While AI can process information and generate outputs at superhuman speeds, the computational resources required — particularly for frontier models with billions of parameters — are extraordinarily expensive. Cloud compute costs, GPU rental rates, and energy consumption for AI inference combine to create a cost structure that, in many enterprise applications, exceeds what companies would pay for human labor performing the same tasks.

This finding has profound implications for the AI industry’s growth narrative. If AI cannot achieve cost parity with human labor for a broad range of tasks, the economic case for widespread AI adoption becomes significantly weaker. Companies that have been investing heavily in AI infrastructure may find that their ROI calculations need substantial revision. The cost challenge is particularly acute for companies attempting to deploy AI at scale, where inference costs compound rapidly. Until compute efficiency improves dramatically or model architectures become significantly more efficient, the promise of AI-driven cost savings may remain largely theoretical for many enterprise applications.

5. Google Debuts New AI Models and Personal AI Agents in Effort to Keep Pace with OpenAI and Anthropic

Google has unveiled a new generation of AI models and personal AI agents designed to close the competitive gap with OpenAI and Anthropic, according to reporting from CNBC and the Wall Street Journal. The announcements represent Google’s most significant AI product push in months and signal the company’s determination to maintain its position in the rapidly consolidating frontier AI market.

The new models and agents span multiple product lines, including personal AI assistants that can handle complex, multi-step tasks across Google’s ecosystem. According to The Tech Buzz and blog.google, the updates include a new era for AI Search, the introduction of “Ask YouTube” — an AI-powered conversational search feature for video content — and the integration of Gemini Omni into YouTube Shorts. The personal AI agents are designed to operate autonomously across Google’s services, from Gmail to Calendar to Maps, creating a more cohesive and intelligent user experience.

The competitive context for these launches is critical. OpenAI and Anthropic have both been gaining ground rapidly, with Anthropic’s Claude models earning particularly strong reviews for reasoning capabilities and safety, while OpenAI continues to dominate the consumer AI chatbot market. Google’s Gemini, despite the company’s vast resources and data advantages, has struggled to achieve the same level of user enthusiasm. The new releases represent Google’s attempt to leverage its ecosystem advantages — billions of users across Search, YouTube, Gmail, and Android — to create AI experiences that competitors cannot easily replicate. Whether these launches will be sufficient to shift market dynamics remains to be seen, but they demonstrate that the AI competition is far from settled.

6. NVIDIA Expands Open Model Families to Power the Next Wave of Agentic, Physical and Healthcare AI

NVIDIA has announced a major expansion of its open model families, targeting agentic AI, physical AI, and healthcare applications, according to the company’s official newsroom. The expansion represents one of NVIDIA’s most significant commitments to open-source AI and reflects the company’s strategic vision for the next phase of AI development — one that extends beyond text and images into robotics, scientific discovery, and autonomous systems.

The announcement, covered by The Next Platform, IT Pro, and Yahoo Finance, includes new model families designed specifically for agentic AI applications (AI systems that can act autonomously to accomplish goals), physical AI (AI embedded in robots and other physical systems), and healthcare AI (models optimized for medical imaging, drug discovery, and clinical decision support). NVIDIA’s approach to open models is distinctive: rather than releasing individual models, the company is building entire model families with standardized architectures that developers can customize for specific applications. The Next Platform describes this as “the open agentic AI world according to Nvidia” — a vision where NVIDIA’s hardware and software stack becomes the foundational infrastructure for the next generation of AI applications.

The expansion also includes new model coalitions and agentic frameworks designed to promote interoperability across the AI ecosystem, according to IT Pro. This is significant because the current AI landscape is highly fragmented, with models from different labs often incompatible with each other’s tooling and infrastructure. By promoting open standards and interoperability, NVIDIA is positioning itself as the connective tissue of the AI industry — a role that reinforces its dominance in GPU hardware while expanding its influence into the software layer. The announcement comes alongside news of a collaboration with Google Cloud to advance agentic and physical AI, per HPCwire, further cementing NVIDIA’s position at the center of the AI infrastructure ecosystem.

7. Anthropic and OpenAI Launch AI Services Companies in India — Threatening TCS, Infosys, and Wipro

In a bold strategic move, both Anthropic and OpenAI have launched AI services companies in India, directly challenging the country’s dominant IT services giants — TCS, Infosys, and Wipro. According to India Today, The Economic Times, Business Standard, and NDTV Profit, the launches represent a fundamental shift in how frontier AI companies are approaching the enterprise market and could reshape India’s $250 billion IT services industry.

The move by Anthropic and OpenAI is particularly significant because these companies have historically focused on building AI models and platforms, not on providing enterprise services. By establishing dedicated services operations in India, they are effectively competing with the very companies that have been their largest enterprise distribution partners. Jefferies analysis, reported by NDTV Profit, suggests that Indian IT giants should indeed be worried: the frontier AI labs possess technical capabilities and brand recognition that traditional IT services companies cannot easily match, and their direct entry into the services market could bypass the intermediaries that have historically controlled enterprise AI adoption.

The implications for India’s IT sector are profound. TCS, Infosys, and Wipro have built their businesses on being the bridge between cutting-edge technology and enterprise implementation. If Anthropic and OpenAI can provide that bridge directly — offering both the AI technology and the implementation services — the value proposition of traditional IT services companies is significantly diminished. Analytics India Magazine reports that “Indian IT’s biggest AI partners are also its newest enterprise threats,” capturing the paradox at the heart of this development. For Indian IT workers, the competitive pressure could accelerate both job displacement and the need for reskilling, as the nature of AI services work shifts from implementation to strategic consulting and custom development.

📊 Trend Watch

🔭 What to Watch

• Vatican-AI Encyclical Impact — Pope Leo’s encyclical on human dignity and AI, developed alongside Anthropic, could reshape how the world’s 1.3 billion Catholics and many other faith communities think about AI ethics. Watch for how tech companies respond to this moral framing and whether it influences regulatory debates in Catholic-majority countries.

• AI Stock Correction Risk — With BofA warning of a railroad-era-level bubble and the S&P 500 near 7,500, any negative catalyst — disappointing earnings from AI-heavy companies, compute cost revelations, or regulatory shifts — could trigger a significant market correction. The concentration of market cap in AI-related stocks amplifies systemic risk.

• Enterprise AI Cost Reckoning — Microsoft’s data showing AI costs exceed human labor costs is likely just the tip of the iceberg. As more companies publish their AI ROI figures, expect a wave of reassessments about which AI applications are economically viable and which remain experimental.

• India IT Disruption — Anthropic and OpenAI’s entry into India’s services market could trigger a rapid consolidation in the Indian IT sector. Watch for TCS, Infosys, and Wipro’s strategic responses — they may accelerate their own AI model development or seek partnerships with alternative AI providers.

• Google’s AI Competitive Position — Google’s new model and agent launches will be judged by market reception and developer adoption. If these releases fail to close the gap with OpenAI and Anthropic, Google may face increasing pressure from investors to restructure its AI strategy or consider acquisitions.