anonym.legal

By · Last updated 2026-04-16

Povratak na blogBezbednost veštačke inteligencije

Nakon incidenta sa 900K korisnika ekstenzija

U januaru 2026. dva maliciozna Chrome dodatka instalirana kod 900K+ korisnika eksfiltrirala su kompletne ChatGPT i DeepSeek razgovore svakih 30 minuta.

April 16, 20268 min čitanja
malicious Chrome extensionAI extension security auditextension trust verificationlocal processing architecture900K extension incident

Incident u januaru 2026.

Azurirano za 2026. U januaru 2026. otkrivena su dva maliciozna Chrome dodatka sa 900.000+ korisnika.

Njihova imena izgledala su kao pravi AI alati:

  • "Chat GPT for Chrome with GPT-5, Claude Sonnet and DeepSeek AI" - 600.000+ korisnika
  • "AI Sidebar with Deepseek, ChatGPT, Claude and more" - 300.000+ korisnika

Oba su radila isto. Na svakih 30 minuta slala su kompletne AI razgovore na udaljeni server. Ukradeni podaci ukljucivali su kod, licne informacije, pravne beleske i poslovne planove. Astrix Security je ovo potvrdio.

Ovi dodaci su trazili dozvolu za "prikupljanje anoniminih, ne-identifikujucih analitickih podataka". Ta formulacija zvuci bezbedno. Nije bila. Prikupljeni podaci bili su potpuno identifikujuci i veoma osetljivi.

Problem bezbednosne inverzije

Korisnici koji instaliraju AI alate za privatnost zele zastitu. Slucaj iz januara 2026. pokazuje najgori ishod: alat koji ste instalirali radi privatnosti je taj koji krade vase podatke.

Ovo nije teorija. Desilo se 900.000 korisnika odjednom. Skeniranje Chrome Web Store-a to nije otkrilo. Korisnicke recenzije to nisu otkrile. Krada je bila skrivena kao "analitika".

Incogni je pronasao da 67% AI Chrome dodataka aktivno prikuplja korisnicke podatke. Za IT timove, kljucno pitanje nije "da li ovo prikuplja neke podatke?" Vec: "mogu li da proverim da ovaj dodatak ne moze da salje sadrzaj razgovora trecoj strani?"

Test provere arhitekture

Postoji jedna pouzdana provera lokalnog procesiranja: pracenje mreze.

Dodatak koji detektuje PII lokalno proizvodi nultu odlaznu saobracaj tokom detekcije. Nijedna veza ka eksternom serveru ne sme se pojaviti izmedju korisnikove paste i slanja ka AI platformi. Samo obradjeni prompt ide van.

Dodatak koji usmerava saobracaj kroz proksi salje vas sadrzaj na server trece strane. Taj operater servera sada je deo vaseg modela pretnji.

Koraci IT provere su jednostavni:

  1. Deploujte dodatak u nadziranoj mrezi
  2. Pokrenite test promptove
  3. Proverite odlazne veze ka serverima izdavaca tokom PII obrade

Ako ne prodje ovaj test, ne odobravajte ga. Marketing tvrdnje nisu bitne. Mrezni saobracaj je dokaz.

Lokalno procesiranje je pouzdano jer je proverljivo. Ne morate da verujete izdavacu. Mozete direktno posmatrati ponasanje. Pogledajte kako anonym.legal ovo resava u nasem pregledu bezbednosti Chrome ekstenzije i vodicom za uskladjenost.

Sta IT timovi treba da zahtevaju

Nakon januara 2026. letvica za AI alate za pregledac mora biti visa.

Minimalna lista:

  • Lokalno procesiranje - provereno mreznom revizijom, ne samo tvrdnjom
  • Poznati izdavac - realna kompanija, jasan poslovni model
  • Nezavisna sertifikacija - ISO 27001 ili ekvivalent
  • Bez usmeravanja kroz server developera za kljucne funkcije privatnosti

Vecina AI dodataka za pregledac nece proci ovu listu. Stopa prikupljanja od 67% to jasno pokazuje. Veliki broj instalacija nije signal bezbednosti. Alati iz januara 2026. imali su stotine hiljada korisnika pre nego sto je iko proverio.

Za vise informacija o bezbednim AI alatima za pregledac, pogledajte nasu stranicu bezbednosti i uskladjenosti.

Izvori

Povezani članci

Bezbednost veštačke inteligencije

AI Coding Assistants Leak Production PII

Unit test fixtures with real customer records. Log files with production data for debugging. GitHub found 39 million secrets leaked in 2024.

Bezbednost veštačke inteligencije

Internal Wiki PII: Confluence Customer Data

Support teams document processes with screenshots of customer accounts. Over 3 years, that's thousands of GDPR data minimization violations in your.

Bezbednost veštačke inteligencije

Screenshot PII: Leaks in Internal Tools

Slack, Teams, Jira, and email regularly receive screenshots containing customer PII. This access-control violation bypasses every DLP tool.

Spremni da zaštitite svoje podatke?

Počnite sa anonimizacijom PII sa 285+ tipova entiteta na 48 jezika.

Započnite besplatnu probuPogledajte funkcije

About this page

We update this page when our platform or the law changes.

Read our founder note for how we work.

Each change shows up in the timestamp at the top.

Related reading

We follow these rules

  • GDPR (EU 2016/679).
  • ISO/IEC 27001:2022.
  • NIS2 (EU 2022/2555).
  • HIPAA safe harbor under 45 CFR § 164.514(b)(2).

Our promise

We do not sell your data.

We do not train models on your text.

We store your files in Germany.

You can delete your account at any time.

You own your work.

Where we run

Our servers live in Falkenstein, Germany.

We use Hetzner. They hold ISO 27001 certification.

All data stays in the EU.

Backups run every day.

Need help?

Email support@anonym.legal.

We reply within one business day.

How we test

We run a full check suite on every release.

Each surface gets its own sweep script and report.

Human reviewers spot-check the output each week.

We track recall and precision on a labelled set.

Bad runs block the deploy.

What we never do

  • We never sell your information to third parties.
  • We never train models on what you upload.
  • We never keep your work after you delete it.
  • We never share keys with any outside firm.
  • We never run ads inside the product.

Plans in plain words

We sell credits, not seats.

One credit covers one short job.

Long jobs use a few credits each.

You can top up at any time.

Unused credits roll over each month.

Read the plans page for current rates.

Who built this

A small team of engineers and lawyers built this.

We ship from Europe and work in the open.

Our founder note spells out why we started.

Where to start

How the parts fit

A browser add-on cleans text inside Chrome.

A Word plug-in handles drafts in Office.

A small desktop tool works on whole folders.

An agent protocol link feeds large models safely.

All four share one core engine and one rule set.

Words from our team

We started this work after a lunch about cookies.

One friend kept getting odd ads on her phone.

We asked why a court file leaked through a draft.

We sketched the first build on a napkin that week.

By month three we had a tiny demo for a friend.

She used it on her first case the next day.

Common questions we hear

Can the tool read scanned PDFs? Yes, with OCR.

Does it work on long files? Yes, in small chunks.

Can I roll my own rule set? Yes, save it as a preset.

Does it run offline? The desktop build runs offline.

Do you keep my files? No, the cloud build wipes after each run.

Will it learn from my work? No, we never train on inputs.

A short tour of the workflow

Upload a file or paste a snippet of prose.

Pick the entities you want gone from the draft.

Choose a method: replace, mask, hash, encrypt, or redact.

Press run and watch the side panel show each hit.

Skim the result and tweak any rule that misfired.

Save the cleaned file or send it to a teammate.