Press & Media
Wzdom tracks AI privacy in real time — and publishes independent research on how major AI companies handle your data.
Wzdom builds privacy-first software for consumers and publishes research on AI privacy, AI ethics, and digital rights. Everything runs on your device. Nothing leaves it.
Wzdom at a glance
Latest from Wzdom Research
Company overview
Wzdom Privacy is a privacy monitoring application for macOS and Windows that intercepts and classifies every network connection a computer makes — trackers, data brokers, advertising platforms, phishing domains — and presents the information in plain English, in real time.
Everything runs locally on the user's device. There is no cloud component, no user account, and no telemetry. The encryption key for the local database is derived on-device and never transmitted. The architecture is privacy-by-design: we cannot access user data because we have no mechanism to receive it.
Wzdom also publishes independent research through Wzdom Labs, covering AI privacy practices, digital rights, and the ethics of AI systems. Research is based on publicly available policies, regulatory filings, and academic sources.
Wzdom Privacy is currently available free during public beta. The company is Wzdom, LLC, founded in 2026 and headquartered in the United States.
Quotable findings — Wzdom Research
All figures from the Wzdom AI Privacy Index (WAPI), Q2 2026. Based on 12 major AI tools analyzed. Full paper ↗
90% of AI tools store your conversations on remote servers — only fully local tools collect nothing.
70% of AI tools use consumer data to train their models in some form — either by default or via limited opt-out.
50% train on user data by default. Users must actively find and change a setting to stop it — if an opt-out exists at all.
30% of AI tools offer no credible opt-out from training whatsoever. Meta AI and DeepSeek have no opt-out; Grok's is buried inside unrelated X account settings.
60% share data with third parties or advertising partners — including ad targeting networks, parent-company cross-use, and sub-processor chains.
60% allow human employees or contractors to read sampled conversations for quality review — a practice rarely disclosed at point of use.
40% of tools scored D or F in the Wzdom AI Privacy Index — meaning fundamental user privacy protections are absent.
Only 10% of tools operate with zero cloud processing. Ollama is the sole fully local AI tool in our analysis — no server, no account, no telemetry.
Media topics we can comment on
Wzdom is available for expert commentary, quotes, and background briefings on the following topics:
- AI tool privacy policy changes
- AI regulation (EU AI Act, U.S. policy)
- Browser extension privacy risks
- AI ethics & accountability
- Consumer data collection practices
- Digital sovereignty & local-first AI
- Data broker ecosystems
- Network-level privacy monitoring
What Wzdom Research covers
Wzdom Labs publishes independent, ongoing research across two domains — AI privacy and AI ethics. Both are covered without commercial affiliation to any of the tools we score or analyze.
- AI privacy practices of major consumer AI tools
- AI ethics, accountability, and explainability
- Data collection by AI platforms — defaults, opt-outs, and hidden practices
- Digital rights and sovereignty
- Regulatory developments — EU AI Act, U.S. policy, global jurisdictions
- Local-first AI — architecture, implications, and adoption
Research standards
Wzdom Research bases all findings on independently verifiable sources. Our methodology:
- Publicly available privacy policies and terms of service
- Legal filings and regulatory enforcement actions
- Academic research and peer-reviewed analysis
- Official regulatory guidance (EU DPAs, FTC, ICO, and others)
- Verified updates — findings are reviewed and updated as policies change
All research is updated on a rolling basis. The WAPI edition, dates, and methodology are published inline with every report. Journalists may cite findings with attribution to Wzdom Research.
From Wzdom
"AI is becoming part of everyday computing — but most people still have little visibility into how these systems collect or use their data. Wzdom exists to make that visible, and to give users real tools to take back control. Privacy should not require a law degree or a subscription."— Founder, Wzdom
Brand assets
Download assets directly below. For SVG logo variants, high-resolution layered files, or anything not listed here, email press@wzdom.ai.
Brand guidelines
The following guidelines cover the minimum requirements for accurate, on-brand representation of Wzdom in press and editorial contexts. For questions or edge cases, contact press@wzdom.ai.
The amber + cream on deep dark combination is a deliberate departure from the blue and green palettes standard across the AI and security industry. Do not substitute colours without approval.
- UI & body: Inter (weights 400, 500, 600, 700, 800). Available free via Google Fonts.
- Editorial headings: Playfair Display (weights 700, 800; italic variants). Used for hero headings and pull quotes.
- Do not use Helvetica Neue, SF Pro, Futura, or their derivatives as substitutes — these undermine the distinctive character of the Wzdom brand.
- Primary usage: Cream wordmark on deep dark (#0c0d18) background, with amber as the single accent on the mark.
- Reversed: Deep dark wordmark on cream (#faf7f0) background.
- Minimum display size: 80px wide. Do not display the logo at sizes where the wordmark becomes illegible.
- Clear space: maintain a margin equal to the height of the "W" mark on all four sides.
- Do not recolour, distort, rotate, apply drop shadows, add outlines, or modify the mark or wordmark in any way.
- Do not place the logo on backgrounds that reduce contrast below WCAG AA requirements.
When writing about Wzdom, use language that is:
- Confident, not arrogant. Wzdom makes strong claims because the evidence supports them — not for marketing effect.
- Warm, not soft. Privacy and AI are serious topics. Wzdom treats users as intelligent adults.
- Technically credible, not jargon-heavy. Explain what something does before naming the technology.
- Honest, not preachy. Wzdom publishes findings because users deserve them — not to moralize.
Online presence
Working on an AI privacy story?
We can provide expert commentary, data, background briefings, and on-the-record quotes. Wzdom Research findings are available for citation with attribution.
press@wzdom.aiWe typically respond within one business day.