// Ethics

Source: The Register

Companies that bet on proprietary AI platforms are discovering they cannot switch vendors as cheaply or quickly as they assumed. Retraining costs, data migration friction, and organizational inertia mean competitive model swaps take months instead of weeks. This vendor stickiness mirrors cloud infrastructure lock-in but operates faster—models deprecate and new competitors emerge rapidly, yet enterprises remain trapped in expensive integration debt. The cost isn't switching itself; it's the competitive disadvantage of being forced to stick with yesterday's model while competitors run newer, cheaper alternatives.

Source: Machinesociety

A controlled experiment by Tan and Nov at NYU Tandon asked 240 adults to interact with chatbots that either showed their "thinking" or simply output answers, probing whether visible deliberation—even if mechanically generated—makes users trust AI systems more. The research isolates a design question: does showing reasoning steps (chain-of-thought outputs, step-by-step breakdowns) increase user confidence independent of actual accuracy, and if so, should it? The gap matters: people want to see the work, but visible reasoning doesn't necessarily correlate with reliability.

Source: WIRED

ChatGPT and similar models lack fiduciary responsibility, real-time market data, and the ability to understand individual tax situations or long-term financial goals—yet people are already using them as free alternatives to paid advisors. A plausible-sounding but incorrect recommendation on tax strategy or asset allocation could cost someone thousands in lost gains or penalties, with no recourse. This exposes a gap between AI capability marketing and actual reliability. Regulators and users now face a practical question: whether "good enough" guidance from a machine is acceptable when real money is at stake.

Source: Seth's Blog

Seth Godin identifies a specific risk: as AI becomes ubiquitous, organizations use it as cover for avoiding hard work. The deflection is straightforward—blame the technology, the transition period, unpredictability—rather than make difficult choices about quality, service, or innovation. The timing matters. "We're still figuring out AI" stopped being credible some time ago. It now signals that leadership either lacks conviction or has decided to coast.

Source: Axios

The Catholic Church is inserting itself into AI governance conversations by weaponizing its moral authority at a moment when Silicon Valley and national governments have largely failed to establish enforceable ethical frameworks. By framing AI regulation through Catholic social teaching and papal authority rather than technical standards or legislation, the Vatican is creating a parallel institutional track that could influence how billions of Catholics—and their governments—approach AI deployment in healthcare, education, and media. Tech companies want self-regulation, governments want national control, and the Church wants a seat at the table by offering something neither can claim: a centuries-old institution with explicit moral doctrine.

Source: The New Yorker

Schools are adopting AI tools at scale without evidence they improve learning outcomes, driven by vendor marketing and administrative convenience rather than pedagogical need. The core constraint is that educators lack institutional power to resist adoption decisions made by district IT departments and vendors positioning AI as inevitable infrastructure. Until schools develop gatekeeping capacity and demand proof of efficacy before deployment, AI integration will remain a technology-first phenomenon where teachers bear the burden of making tools designed for extraction and optimization serve learning.

Source: The New Yorker Daily

Schools are experiencing delayed reckoning with AI adoption. Healthcare, dating apps, and content platforms embedded the technology before serious pushback emerged. Education's resistance reflects a specific vulnerability: AI's opacity and hallucination risk pose direct threats to knowledge transmission and credentialing—the two functions schools actually protect. What's at stake isn't efficiency or personalization, but whether schools can maintain their role as arbiters of what's true and verifiable when AI systems have become unreliable information sources.