Skip to content
The Crash Log
AI & Tech Gone Off the Rails
Fund
Cover image for The Crash Log newsletter
Issue #023 · May 28, 2026

Cuts Without Returns

A Gartner study says AI layoffs don't pay back, while the pope calls them a social calamity.

STACK_OVERFLOW

80% of Companies Piloting AI Cut Workers; Gartner Says Cuts Aren't Paying Off

A Gartner survey of 350 global executives at companies with at least $1 billion in revenue found that roughly 80 percent of organizations piloting or deploying autonomous AI capabilities reported workforce reductions. The same study found that those reductions did not correlate with stronger return on investment, with companies cutting staff and companies amplifying staff reporting nearly identical AI ROI outcomes. (Source: Gartner)

Helen Poitevin, a VP analyst at Gartner, said in the release that the impulse to demonstrate AI returns through headcount cuts is "misplaced." Organizations seeing real ROI gains, she said, are the ones investing more in human skills and operating models rather than eliminating roles. (Source: Fortune)

The data is landing in the middle of a wave. Tech-sector layoffs in 2026 have already passed 100,000 jobs, with roughly 50,000 of those explicitly tied to AI restructuring. (Source: TechSpot) In May alone, Intuit cut 3,000 workers — 17 percent of its staff — while citing an AI focus shift, Meta began laying off 8,000 workers under the same framing, Coinbase eliminated roughly 700 employees, and PayPal announced plans to eliminate about 4,760 roles, or 20 percent of its workforce, over the next two to three years. (Source: Yahoo Finance)

RUNTIME_ERROR

Pope Leo XIV Releases 42,000-Word Encyclical Warning AI Could Trigger "Social Calamity"

Pope Leo XIV released his first encyclical on Monday — a 42,300-word document titled "Magnifica humanitas: On Safeguarding the Human Person in the Time of Artificial Intelligence." The encyclical was signed May 15, the 135th anniversary of Leo XIII's Rerum novarum, the 1891 text that anchored Catholic teaching on labor and capital. (Source: Vatican News)

The document warns against "a race for ever more powerful algorithms and larger datasets" driven by "the desire to secure geopolitical or commercial dominance," and calls for stronger regulation of private AI companies, worker protections, and constraints on autonomous weapons systems. The pope frames mass unemployment from AI adoption as "a true social calamity that especially requires the State to exercise responsibility," and writes that "the pursuit of greater profits cannot justify choices that systematically sacrifice jobs, because the human person is an end, not a means." (Source: TIME)

The Vatican presented the encyclical alongside Anthropic co-founder Chris Olah — a deliberate signal about which AI lab the Holy See views as a credible partner on safety, given that Google and OpenAI were not on the stage. (Source: Anthropic) Vice President JD Vance called the warnings "profound."

The encyclical's structure echoes Rerum novarum, Centesimus annus, and Laudato si', explicitly placing AI inside the Church's labor and ecology tradition. (Source: Axios)

Stack Trace

Anthropic said Tuesday that its Project Glasswing partners — including Cloudflare, Cisco, Oracle, and Microsoft — used Claude Mythos Preview to identify more than 10,000 high- or critical-severity vulnerabilities in critical software in one month. Cloudflare alone found 2,000 bugs across its critical-path systems with a false-positive rate the company says beats human testers. (Source: Help Net Security) Anthropic still won't release Mythos publicly — the company says no one, including itself, has built safeguards strong enough to prevent the same model from being weaponized for offensive use. (Source: The Hacker News)

Pennsylvania Governor Josh Shapiro's administration sued Character Technologies, alleging Character.AI chatbots illegally hold themselves out as licensed medical professionals — in violation of the state's Medical Practice Act. (Source: NPR) A state investigator interacting with a bot named "Emilie" was told it had attended medical school at Imperial College London, was licensed to practice psychiatry in Pennsylvania, and could prescribe medication. The bot supplied a fabricated Pennsylvania medical license number. (Source: The Philadelphia Inquirer)

OpenAI announced at Google I/O 2026 that every image generated through ChatGPT, the OpenAI API, or Codex since May 19 carries two invisible markers — C2PA provenance metadata plus Google DeepMind's SynthID pixel-level watermark — that survive screenshots, compression, and format conversion. (Source: PhotoWorkout) A public verifier tool at openai.com/verify lets anyone test an image. Caveat: it only catches OpenAI-generated images; Midjourney, Stable Diffusion, and the open-source pipeline behind most political deepfakes remain unaddressed. (Source: IBTimes)

Don't miss the next issue

Subscribe