ai ethics - Adjacent

Oct 7

Sam Altman's Dangerous Comfort With AI Collateral Damage

Altman's framing—that society must tolerate "bad things" as the price of AI advancement and personal agency—inverts the actual power dynamic. A handful of companies are centralizing control over infrastructure while externalizing harms onto the public, not the reverse. This rhetorical move preempts regulation and accountability by repackaging corporate risk-taking as inevitable technological progress, a playbook that has already failed with social media, algorithmic finance, and surveillance capitalism. OpenAI is willing to cause harm, yes—but the deeper point is they believe they've won the narrative enough to say it out loud.

theme-culture regulation/policy ai ethics healthcare

Sep 20

Doctors Push Back on Medical AI Beyond Imaging and Diagnostics

Source: Financial Times (paywall)

Clinicians are drawing a hard line between AI applications they trust—image analysis where performance metrics are clear and historical data abundant—and broader clinical deployments where algorithms influence treatment decisions without comparable evidence. The distinction reflects a real stakes difference: misdiagnosing a chest X-ray affects one patient; a flawed AI system recommending drug dosages or patient triage affects workflows and liability across entire hospital systems. Venture-backed medical AI companies will face a credibility wall without prospective clinical trial data that regulators and providers increasingly demand before deployment.

theme-culture regulation/policy ai ethics IP Rights

Sep 18

Microsoft Exec Condemns AI Training Data as "Largest Labor Theft"

Source: Ars Technica

A Microsoft executive's invocation of "theft" rhetoric signals how AI copyright disputes are shifting from technical and legal arguments toward moral framings that resonate with workers worried about displacement. Microsoft itself has aggressively scraped data to train its own models, which undercuts the company's credibility. But the framing is now in circulation, giving plaintiffs and critics a resonant language that courts and regulators may adopt. AI companies can no longer control the narrative solely through IP law; the debate is becoming cultural and political.

theme-culture regulation/policy ai ethics autonomous systems

Sep 17

AI's Speed in Warfare Multiplies the Risk of Killing Mistakes

Source: Financial Times (paywall)

The Financial Times examines how military adoption of AI for targeting—prized for its velocity and capacity to process vastly more data than human analysts—introduces a new class of systematic error at scale. Faster target identification saves lives in some scenarios but creates cascading failure modes in others, where algorithmic mistakes can propagate across entire strike campaigns before human review catches them. The problem is structural: the speed advantage that makes AI militarily attractive also embeds error at a scale human oversight cannot match in real time.

theme-culture regulation/policy ai ethics legal system

Sep 12

Lawyer Fined for Using AI-Generated Witnesses in Murder Defense

Source: The Verge

A New Mexico court fined a defense attorney $5,000 for submitting AI-hallucinated evidence, the first major judicial penalty for courtroom AI failures. The ruling establishes that generative AI's tendency to invent plausible-sounding details carries real legal liability. Lawyers and legal tech vendors now face pressure to close the gap between LLMs' fluency—which sounds authoritative—and their unreliability on factual claims. This could accelerate adoption of verification-first legal AI tools or trigger stricter gatekeeping over what can be filed in court. Judges are holding individual practitioners accountable rather than treating AI errors as technical inevitabilities. That approach creates a split incentive: overworked public defenders may avoid AI assistance entirely, while well-resourced firms build compliance infrastructure.

theme-culture ai ethics animal welfare regulation/policy

Sep 6

AI translators for animal communication raise exploitation risks

Source: Bloomberg (paywall)

Researchers are using machine learning to decode animal vocalizations and behaviors with new precision. Bioethicists warn of a concrete risk: the ability to understand animal communication creates new avenues for industrial-scale manipulation—factory farming optimization, wildlife trafficking networks using decoded distress signals. The asymmetry is stark. Human translation involves mutual speakers. Decoding animal communication is extractive: it gives humans a one-way window into creatures that cannot consent to being understood or influenced at scale.

theme-culture regulation/policy labor ai ethics

Aug 31

Robotaxis Displace Thousands of Professional Drivers

Source: TechCrunch

As autonomous vehicle deployments accelerate in San Francisco and Phoenix, the industry is largely ignoring the immediate economic impact on the 3.5 million professional drivers in the U.S., many of whom lack the education or transferable skills for retraining. The robotaxi narrative—framed as inevitable progress—obscures labor displacement happening now, not in some distant future, with minimal policy discussion about transition support, pension protection, or alternative livelihoods for a vulnerable workforce already operating on razor-thin margins.

theme-culture media regulation ai ethics

Aug 30

Musicians Are Becoming AI Fraud Investigators

Source: The Verge

A growing cohort of artists with deep technical knowledge of production are now actively identifying and exposing AI-generated music passing as human work. This crowdsourced authentication undercuts the idea that detection will remain an automated, algorithmic problem. Rather than platforms or copyright holders controlling the narrative, individual creators with skin in the game are becoming the arbiters of authenticity. This creates both more granular accountability and new power dynamics around who gets to declare something "real." As generative tools democratize, so does the labor of resistance. Cultural gatekeeping is shifting from institutional authority toward community expertise.

theme-culture regulation/policy ai ethics content rights

Aug 15

OpenAI's Crawler Ignores Robots.txt Rules for Training Data

Source: Search Engine Journal

OpenAI is allowing its GPTBot to bypass robots.txt directives that publishers use to prevent automated access, treating the standard as advisory rather than binding. This escalates friction between AI labs and content creators. Publishers lack technical recourse to stop training scrapes and must rely on legal action, shifting power to AI infrastructure companies that can unilaterally decide which rules apply to them.

theme-culture regulation/policy media ai ethics

Aug 14

Self-Represented Litigant Embeds AI Manipulation Code in Court Filing

Source: 404 Media

A Connecticut pro se defendant inserted prompt injection attacks into their legal filing, attempting to hijack any AI system a judge might use to summarize or analyze the case. This is the first documented instance of adversarial prompt engineering deployed in actual litigation. The attack exposes a gap: courts are adopting AI to manage document volume faster than they are securing it.

theme-culture regulation/policy ai ethics education

Aug 4

Mexico's largest university scraps AI exam proctoring after mass failure

Source: Ars Technica

UNAM's decision to invalidate 58,000 exam results—roughly a third of test-takers—exposes the gap between AI proctoring vendors' claims and classroom reality, particularly in resource-constrained settings where internet reliability and device access are uneven. The failure forced a complete retake, signaling that universities betting on remote proctoring as a cost-saver are underestimating both the technical and social risks. The incident will likely slow adoption of AI exam supervision in Latin America and reinforce skepticism among administrators already wary of outsourcing assessment to opaque systems.

theme-culture regulation/policy education ai ethics

Jul 26

Elite universities abandon AI detection tools over accuracy failures

Source: Financial Times (paywall)

Yale, Johns Hopkins, and Waterloo rejected AI detectors after the tools produced enough false positives to damage student grades and academic standing. The unreliability exposed a fundamental mismatch: universities want instant detection, but need reliable assessment. As institutions retreat, the work reverts to human review. AI detection will remain a supplementary flag in education, not a basis for enforcement decisions.