AI TRUST

Who is the chatbot working for?

LessBother tracks commercial pressure, advertising, privacy rules, tracking practices and major policy changes across consumer AI products.

No mystery score yet.

This reviewed baseline was refreshed August 29, 2026. The live watcher below preserves later wording changes automatically. Numeric trust scores will only appear after the methodology has enough historical evidence to justify them.

Checking whether any monitored source changed after this reviewed baseline…

AUTOMATIC SOURCE WATCH

The policy pages watch themselves.

LessBother checks official privacy, advertising, commerce and data-control pages on a schedule. It stores a fingerprint of the relevant wording and records a before/after change when that wording moves.

Watcher tags are topic labels only. A tag such as “advertising” or “tracking” means that topic appeared in changed wording. It does not mean the policy became better, worse, more invasive or more commercial until the evidence is reviewed.

Loading source-watch status…

Recent detected changes

Loading change history…

AUDIT THIS RECOMMENDATION

What did the chatbot just steer you toward?

Paste a shopping or product recommendation from any AI. LessBother checks the text in this browser for brands and exact products already in our evidence catalog, then points you to what we actually know.

This does not prove paid steering. A brand mention is only a lead to inspect. LessBother does not upload or save the pasted answer.

ChatGPT

OpenAI

Consumer ads: Ads are documented for Free and Go users in supported markets; OpenAI says ads do not change ChatGPT answers.
Advertising stack: OpenAI publishes Advertising Terms plus optional Ad Tools covering conversions and audiences.
Ad data: OpenAI's U.S. privacy policy says Free/Go ad personalization and ad-effectiveness measurement can use personal data subject to settings and applicable controls.
Model improvement: Consumer content may be used to improve models unless the user exercises available data controls; separate temporary/private modes have different handling.

Testing ads in ChatGPT ↗OpenAI U.S. Privacy Policy ↗OpenAI Advertising Terms ↗

Open ChatGPT evidence profile →

Claude

Anthropic

Advertising: Anthropic says Claude will remain ad-free and will not include advertiser-influenced responses or unrequested third-party product placements.
Commerce: Anthropic says it is interested in user-initiated agentic commerce and product/business integrations while keeping the user, not an advertiser, as the initiator.
Model improvement: Anthropic says consumer Inputs and Outputs may be used to train/improve models unless the user opts out, with documented safety/feedback exceptions.
Technical data: The privacy policy describes device, browser, IP-derived location, identifiers, usage information and cookies/similar technologies.

Claude is a space to think ↗Anthropic Privacy Policy ↗

Open Claude evidence profile →

Gemini / Google AI

Google

AI ads: Google documents sponsored retailer formats and newer Gemini-built ad formats in AI Mode/Search, including clearly labeled sponsored placements.
Commerce: Google documents merchant feeds, Direct Offers and UCP-powered checkout reaching AI Mode and the Gemini app.
Privacy: Google's privacy policy describes online identifiers, browsing/search activity, location and ad-interaction data used for advertising, with controls in My Ad Center and other settings.
AI training: Google says publicly available information can help train Google AI models including Gemini Apps and Cloud AI capabilities.

Google Privacy Policy ↗Digital advertising and commerce in 2026 ↗New ad formats built with Gemini ↗

Open Gemini / Google AI evidence profile →

Perplexity

Perplexity AI

Advertising: Perplexity documents sponsored follow-up questions and paid media and says advertisers do not influence the content of answers.
Shopping: Perplexity says related product listings are not paid placements, while merchants providing deeper availability/review/specification data can be more likely to be recommended.
AI learning: Perplexity says search queries and feedback are used to improve the experience; users can turn off AI Data Usage to stop search data being used to enhance AI models.
Data collection: Its help center says it collects device and interaction data and stores personal information when an account is created.

Why we're experimenting with advertising ↗What data does Perplexity collect about me? ↗Shop like a Pro / product recommendation FAQ ↗

Open Perplexity evidence profile →

WHAT THE WATCHER TRACKS

The promises matter when they change.

Commercial pressureAds, sponsored placements, merchant feeds, affiliate or checkout relationships.

Privacy & trackingTraining use, ad personalization, identifiers, retention, sharing and user controls.

Policy driftSave old wording, diff new wording, timestamp the change and keep the source.

Recommendation behaviorConsumer recommendation testing stays a separate evidence lane when reliable collection is available.

THE RULE

Observation is not proof of corruption.

A brand appearing more often is a signal to investigate, not evidence that somebody secretly paid for the recommendation. LessBother separates disclosed commercial relationships, observed behavior and proven influence.