Blog — The information-integrity field notes

Read the field.

How AI is rewiring discovery — and what it takes to keep the factual record straight when a model, not a search page, gives the answer.

Fundamentals

Start with the shift.

What Is Generative Engine Optimization (GEO)?

The discipline that replaces SEO when the answer replaces the link: how to be cited, not just ranked.

July 3, 2026 · 7 min read

AEO vs SEO vs GEO: the three layers of visibility in 2026

Three acronyms, three audiences — search engines, voice assistants, generative models. How to tell them apart and manage all three.

July 21, 2026 · 6 min read

How ChatGPT decides what to say about your brand

Inside the retrieval-plus-generation pipeline: why a model repeats an error no human would believe.

August 8, 2026 · 7 min read

The answer gap: when AI knows less than a search engine

The new knowledge gap between what a search engine can surface and what an LLM dares to assert.

August 26, 2026 · 6 min read

Why being cited matters more than being ranked

In a world of single answers, position #1 loses its meaning — citability becomes the metric that counts.

September 13, 2026 · 6 min read

Deep Narrative Intelligence vs. e-reputation: what changes when the answer replaces the link

Social listening tracks what humans say. DNI tracks what models assert. The difference is not cosmetic — it changes the method, the metrics, and the timescale of correction.

October 23, 2026 · 6 min read

Why Deep Narrative Intelligence must be multilingual — and what happens when it isn't

A model doesn't say the same thing about you in English and in German. The errors are different, the sources are different, the corrections must be too. Here is why monolingual DNI is blind.

October 24, 2026 · 6 min read

Deep Narrative Intelligence: a quick orientation for executives

A five-minute read for executives who need to understand what Deep Narrative Intelligence is, why it matters now, and where it fits in their risk and governance map.

October 24, 2026 · 5 min read

Deep Narrative Intelligence vs. fact-checking: why they are not the same discipline

Fact-checking verifies a single claim against a source. Deep Narrative Intelligence measures the aggregate narrative a model holds and corrects it at the root. The difference matters for how you resource both.

October 24, 2026 · 6 min read
AI disinformation

How falsehoods travel.

How a false claim gets laundered into an LLM

The journey of a lie: from an obscure forum to an answer an AI assistant states with total confidence.

July 6, 2026 · 7 min read

Brand hallucinations: anatomy of an AI factual error

Products that don't exist, executives never hired, lawsuits never filed — how these hallucinations start, and how to correct them.

July 24, 2026 · 7 min read

Poisoned SEO: manipulating a model through its web sources

How bad actors inject narratives into models by targeting the sources those models index.

August 11, 2026 · 7 min read

Deepfakes and fabricated quotes: the new reputation attack

When a false statement attributed to an executive spreads faster than the correction.

August 29, 2026 · 7 min read

The DISARM framework applied to AI-generated disinformation

An open taxonomy for naming, classifying and anticipating disinformation techniques in the generative-AI era.

September 16, 2026 · 7 min read

How Deep Narrative Intelligence traces the provenance of AI errors

A false claim from a model isn't groundless — it came from somewhere. DNI maps the chain: which source published the error, which pages replicated it, and which of those the model reads.

October 23, 2026 · 7 min read

Deep Narrative Intelligence in a crisis: when a model starts lying about you in real time

A viral rumour, a regulatory leak, a competitor's smear — within hours, an AI model can start asserting it as fact. Here is how DNI contains the spread before the next training cycle locks it in.

October 24, 2026 · 7 min read

When the model invents you: hallucinations, entity confusion, and how DNI catches what no human published

The most dangerous AI errors aren't copied from a bad source — they're fabricated from statistical patterns. Here is how Deep Narrative Intelligence detects and corrects claims that exist nowhere on the web.

October 24, 2026 · 6 min read

Deep Narrative Intelligence and the anti-disinformation stack: where DNI fits

The anti-disinformation stack has many layers — fact-checking, media monitoring, OSINT, platform moderation. DNI is a new layer that addresses the model-generated dimension of disinformation. Here is where it fits and why it is distinct.

October 24, 2026 · 6 min read

Deep Narrative Intelligence against influence operations: when state actors weaponise AI models

State-backed influence operations have discovered that AI models are a force multiplier — plant a distortion in the source layer, and the model delivers it to every user who asks. DNI is the defensive discipline that detects and counters this.

October 24, 2026 · 7 min read

AI-generated content and AI citation: the recursive risk Deep Narrative Intelligence addresses

When AI generates content that another AI cites as a source, a recursive loop forms where errors compound and amplify. This is the recursive risk that DNI is uniquely positioned to detect and break.

October 24, 2026 · 6 min read
Brand monitoring

See what AI already says.

How to monitor what ChatGPT, Gemini and Perplexity say about you

The method for turning a vague worry — "what does AI say about us?" — into a measurable monitoring programme with actionable alerts.

September 28, 2026 · 7 min read

Building an AI reputation dashboard

The indicators to track week after week to manage your exposure inside model answers.

July 9, 2026 · 6 min read

The 12-point AI presence audit checklist

The protocol we use to map what supported AI models know — and don't know — about an organisation.

July 27, 2026 · 6 min read

What to do when an AI spreads false information about you

The four-step procedure: document, trace the source, correct, verify persistence.

August 14, 2026 · 6 min read

Measuring the gap between verifiable facts and what AI says

A scoring method to quantify, claim by claim, how faithful a model is to reality.

September 1, 2026 · 6 min read

The AI Image Score: how Deep Narrative Intelligence measures what models say about you

Sentiment scores measure tone. The AI Image Score measures factual accuracy — one number per model, per language, tracked over time. Here is how it works and why it changes the conversation.

October 24, 2026 · 6 min read

Deep Narrative Intelligence for executive protection: when a model defames your CEO

A model that tells a journalist your CEO is under investigation — when they are not — is a defamation event at scale. DNI is how organisations detect, trace, and correct model-generated falsehoods about their executives.

October 24, 2026 · 6 min read

Deep Narrative Intelligence as competitive intelligence: what models say about your rivals

If you know what AI models say about your competitors — their strengths, their weaknesses, their distortions — you hold a strategic asset. DNI turns the model layer into a source of competitive intelligence.

October 24, 2026 · 6 min read
By sector

Risk, by industry.

Finance: why AI assistants must cite verified sources

When a fabricated number can influence an investment decision within seconds.

July 12, 2026 · 7 min read

Healthcare: the risk of clinical hallucination in AI assistants

Why information integrity becomes a patient-safety issue.

July 30, 2026 · 6 min read

Public sector: information sovereignty and generative AI

What institutions must demand from models that answer on their behalf, indirectly.

August 17, 2026 · 6 min read

Energy & industry: protecting reputation against AI narratives

Between a polarised public debate and content-hungry models, how to secure the factual narrative.

September 4, 2026 · 6 min read

Defence: information integrity and language models

Why the sector most exposed to disinformation is also the least ready for LLMs.

September 19, 2026 · 6 min read

AI Information Integrity in the Public Sector | SetRecord

How public institutions can assess AI-generated claims, governance needs, data flows and deployment requirements.

October 24, 2026 · 6 min read

Deep Narrative Intelligence in M&A: the hidden due-diligence gap

Before a merger closes, acquirers audit financials, legal, cybersecurity. Almost none audit what AI models already say about the target. Here is why DNI is becoming a standard step in M&A due diligence.

October 24, 2026 · 7 min read

Deep Narrative Intelligence before an IPO: what models say about you on listing day

On listing day, every journalist, analyst and investor will ask an AI about your company. If the model's answer diverges from your prospectus, you have a problem that no roadshow can fix. DNI closes that gap before the bell rings.

October 24, 2026 · 7 min read

Deep Narrative Intelligence for financial institutions: when a model gets your balance sheet wrong

Banks live and die by trust in their numbers. When an AI model invents, distorts or stale-copies financial figures, it creates a risk that neither PR nor compliance was built to catch. DNI is the layer that does.

October 24, 2026 · 7 min read

Deep Narrative Intelligence for pharma: when AI asserts your drug is unsafe

A model that tells a patient your drug is unsafe — based on a retracted study, a stale label, or a conflation with a different molecule — is a public-health risk and a commercial risk. DNI is how pharma measures and corrects it.

October 24, 2026 · 7 min read

Deep Narrative Intelligence for energy and defence: strategic sectors, strategic narratives

Energy and defence companies operate at the intersection of national security, geopolitical perception, and commercial interest. When AI models distort their narrative, the stakes are strategic. DNI is how these sectors measure and protect their model-layer reputation.

October 24, 2026 · 7 min read
Methodology & evidence

The proof behind the claim.

Provenance: tracing the origin of an AI claim

The central concept of our method — and why a correction without provenance doesn't hold.

July 15, 2026 · 6 min read

ClaimReview and structured credibility signals

How semantic markup helps models tell a verified fact from a rumour.

August 2, 2026 · 7 min read

Building an "AI-ready" fact sheet for your brand

The single source document every model should be able to cite without error.

August 20, 2026 · 6 min read

The disinformation risk score: how it is calculated

Seven variables, one number — the mechanics behind our risk indicator.

September 7, 2026 · 6 min read

Correcting an AI narrative: the 5-step method

From alert to persistence check — the complete protocol, step by step.

September 22, 2026 · 6 min read

Deep Narrative Intelligence: the method that corrects what AI says about you

How SetRecord interrogates AI models with and without web search, quantifies the gap to verifiable facts, and traces the provenance of every erroneous claim — before delivering a sourced correction.

October 22, 2026 · 8 min read

Deep Narrative Intelligence and the training cycle: why correction takes six months, not six days

Models update their deep knowledge every 3–6 months. A correction that works in days on the web layer can take months on the training layer. Here is the full timeline — and why patience is a methodological requirement, not a virtue.

October 24, 2026 · 8 min read

With and without web search: why DNI interrogates in both configurations

A model answers differently with web search enabled and disabled. One reveals the latent narrative baked into the training data; the other reveals the live retrieval layer. DNI measures both, because they are two different risks.

October 24, 2026 · 6 min read

How the AI Image Score evolves over time: reading the curve

The AI Image Score is not a static number — it is a curve. Understanding its shape tells you whether your narrative is stable, improving, regressing, or volatile, and whether your correction programme is working.

October 24, 2026 · 6 min read

Inside a correction cycle: from detection to verified persistence

A correction is not a single act — it is a cycle with five stages, each of which can fail. This article walks through the full cycle, from detecting a distortion to verifying that the correction has persisted across models and time.

October 24, 2026 · 7 min read

Provenance data retention in Deep Narrative Intelligence: what we keep, for how long, and why

DNI generates provenance data — the evidence trail behind every measurement and correction. What we keep, for how long, and why is a governance question that this article answers with full transparency.

October 24, 2026 · 6 min read
Governance & law

Who answers for the answer.

The EU AI Act: what companies need to anticipate

Transparency obligations, indirect reputational risk — what actually changes.

July 18, 2026 · 6 min read

Who is liable when an AI defames your company?

Model publisher, source-content host, distribution platform: the liability chain, explained.

August 5, 2026 · 7 min read

Crisis management in the generative-AI era

Why your crisis communications plan must now include language models.

August 23, 2026 · 6 min read

GDPR and AI: what models know — and shouldn't say — about you

Personal data, the right to rectification: how GDPR applies to a model's outputs.

September 10, 2026 · 6 min read

Building an internal AI-reliability policy (playbook)

Roles, review cadence, alert thresholds: the skeleton of a governance model that lasts.

September 25, 2026 · 6 min read

Deep Narrative Intelligence and European Information Integrity

How the European context shapes AI information integrity, data protection, deployment choices and multi-model monitoring.

October 24, 2026 · 7 min read

Deep Narrative Intelligence under GDPR: how SetRecord audits AI models without breaking the law

Interrogating AI models about a person or organisation is a data-processing activity. Doing it without a legal basis is a liability. Here is how SetRecord built DNI to be GDPR-compliant by construction.

October 24, 2026 · 7 min read

Deep Narrative Intelligence as an instrument of European digital sovereignty

If the narratives that AI models hold about European institutions, companies and policies are shaped outside Europe, sovereignty is incomplete. DNI is the instrument that makes the narrative layer observable and correctable from within.

October 24, 2026 · 7 min read

How to report Deep Narrative Intelligence to your board

Boards understand risk, metrics, and trends. This article shows how to translate DNI findings into the language of board reporting — with the three metrics that actually drive decisions.

October 24, 2026 · 6 min read

The six-month synthesis: how to present DNI results to your executive committee

Every six months, the DNI programme produces a synthesis for the executive committee. This is not a data dump — it is a strategic narrative that connects the AI narrative layer to business outcomes. Here is how to structure it.

October 24, 2026 · 7 min read

See what AI says about you

A SetRecord audit checks what supported AI models say against your verifiable facts — within five business days.