By · Last updated 2026-03-12

Til baka á BloggLögfræðiteknik

E-Discovery sakir: Gervigreind villa i thykkju

I Athletics Investment Group gegn Schnitzer Steel (2024) leiddi rangleg thykkja til sakir vid uppgefning. Med gervigreindar verkfaerum sem naed eingongu 22,7% naekvaemi standa lagateymi frammi vid raunverulega aborgun.

March 12, 202610 mín lestur
e-discovery sanctionsredaction liabilityAI redaction precisiondocument reviewlegal technology

title: "E-Discovery sakir: Naer gervigreind thykkja fer of langt" description: "I Athletics Investment Group gegn Schnitzer Steel (2024) leiddi rangleg thykkja til uppgefningarsakir. Med gervigreindar verkfaerum sem naed eingongu 22,7% naekvaemi standa lagateymi frammi vid raunverulega aborgun." category: legal-tech publishedAt: 2026-03-12 tags:

  • e-discovery sakir
  • thykkju aborgun
  • gervigreind thykkju naekvaemi
  • skjalayfirfarning
  • lagatechnology readingTime: 10

Uppfaert 2026

Tvaer leidir sem thykkja bristr

Lagateymi standa frammi vid tvaer bilanaleidir. Badaer skapar raunverulega aborgun.

Van-thykkja afhjupar leidrettileg gogn eda persounlegar uplysingar sem verda ad vera falin. Adilinn afhendir gagnafaerslur sem honum bar -- og oft var skylda -- ad vernda.

Of-thykkja felur staergir sem motmalaaditl a rett til ad sja. Domstolar medfara thetta sem hindrun. Thad er uppgefningarbrot subject to sakir.

Gervigreind verkfaeri sem faer innkoll umfram naekvaemi veldur odru vandanum med honnun. Gervigreind velur sem blakkir 80% af skjali formas thad ad missa neitt. En nidurstadan er annytta. Hann getur einnig vakid domstolssakir.

Badaer bilanaleidir leda til sama stads: domari, skyringu og kostnad.

Schnitzer Steel malid (2024)

2024 malid Athletics Investment Group gegn Schnitzer Steel syna hvernig domstolar medfara rangleg skjalageymslu.

Einn adilinn framleiddi skjol med breida merkingar. Motmalaaditl lystu ythur sig. Domstollinn leit a gogn. Hann fann merkingar gengu lengra en lagunum leyfir.

Nidurstadan: sakir samkvaemt Federal Rule of Civil Procedure 37. Framleidandi adilinn greiddi fyrir galladir ferill.

Slikkar sakir eru ekki nyjar. Domstolar hafa notad thad i ar. Hvad gerir thetta mal ahugavert er timing. Gervigreind-styddar yfirfarning eru nu algengar i deilumalum. Malid vaknar lykilspurningu: hafa lagateymi athugad naekvaemi gervigreind-verkfaeranna sinna adir en notkun thar i framleidning?

Svari skiptir maeli. Verkfaeri med litla naekvaemi mun merkja allt of mikid. Loerfaedurinn sem byggir a nidurstodunni an athugunar ber ahattuna.

Til ad fa full malslysing, sja E-Discovery LLC greiningu a tengdri geymslu.

22,7% naekvaemivandinn

Presidio er opinn PII greinivelin smidur af Microsoft. Hann er notadar mikid i skjalafyrirfarnar verkfaerum. Prof a domstolsskjolum og samningum gefu honum 22,7% naekvaemi.

Naekvaemi maelir hversu oft jaekvaedt flag er rekt. Vid 22,7%, eru um 77 af hverjum 100 flaggum falskaekvaedar. Thau aetlun eru ekki tiltaekni eftir neinni gildandi staedil.

Fyrir e-discovery, er staerdfraedinn bein. 10.000 skjolasett unnin a tha maete mun hafa thorusanir grundlausra merkinga. Framleidandinn stendur frammi vid somu ahattu og Schnitzer Steel stefndi: graad framleidning, domstolsyfirfarning og moglegar sakir.

Thetta tala er fyrir Presidio i sjafgefdu uppstodunni a lagafyrirtaeki efni. Ekki allar gervigreind verkfaeri starfa a thessu stigi. En thetta vel er algengasta opna uppspretta kosta i greininni.

Astaedan er grundvallarl. NLP-kerfi thraenar a almennum texta. Domstolsmal er annarslegt. Hann notar faglegar hugtok, tilvisanasnid og samninga reglur sem veikja fra thjalfunargognum. Verkfaeri sem virkar vel a laekniskjolum getur gert miklu verr a yfirheyrslur-skrifum.

Hvad gervigreind-notkunargogn syna

Her er odur gagnastadur: 27,4% af gervigreind-spjallbod efni er tiltaekt, samkvaemt sjalfstaedri greiningu a fyrirtaekja gervigreind notkun.

Thetta lySir hvad starfsmenn senda vid venjulegt verkefni. Ekki gogn sem thau aetkudu ad deila -- efni innifalid vegna venja eda slysni. Loerfaedar sem nota gervigreind til ad semja braef, fara yfir samninga eda thatta yfirheyrslur senda tiltaekar gogn a netjona gervigreind-verdusins sem hvartrunarahrif vid venjulegt verk.

Naerri thridi hlutur samskipta fela i ser visnagogn, leidrettileg upplysingar eda malsstefnu. Thad efni naed netjona gervigreind-verdusins i notanlegu myndi nema medir komi.

Fyrir lagafyrirtaeki sem athuga gervigreind-ahattu sinar, er 27,4% ekki smaavar mael. Thad er grunnmagn. Naerri thridji hluti gervigreind-notkunar i fyrirtaeki felur i ser efni sem tharf vern.

Abyrgleikakedjan

Of-geymsla og gervigreind gagnaleki skapa serstak en tengda ahattuleidir. Badar byrja med somu aakvordun: nota gervigreind-verkfaeri an naegilegrar mats.

Uppgefningarleidin: Gervigreind merkir efni breitt -> loerfaedur byggir a nidurstodu an staeduskodunar -> framleidning hefur oreastudar merkingar -> motmalaaditl motmaela -> domstoll fer yfir -> sakir.

Gagnaleykileidin: Loerfaedur notar gervigreind fyrir malsverk -> Gervigreind faer leidrettileg samskipti -> gervigreind-verdur sufnir innbrot -> visnagogn afhjupud -> malpractice krafu fylgja.

Upphafspunkturinn er sami i badaum tilfellum. Fyrirtaeki reka gervigreind-verkfaeri an vitundar um hvad tha verkfaeri gerir raunverulega. Engar stjornandi eru settar upp fyrir verkid.

Naekvaemi-fyrst yfirfarning fyrir framleidning

Domstolar spyrja throngt spurningar naer thad fer yfir umdeladar merkingar. Var hvert studdur af forrettindum, trundlurreglur eda domstolsskipun? Domstolar spyrja ekki hvort verkfaeri framleidandans flaggadi sem magnaed sem moguleg.

Merking an retts grundvalls er uppgefningarbrot. Skiptir ekki maeli hvort mannmaer eda gervigreind gerdhi thad. Rannsokn er merking-fyrir-merking.

Fyrir loerfaeda thythir thetta ad gervigreind-yfirfarnarverkfaeri tharf ad vera prufa a naekvaemi -- hlutfall flagga sem eru raunverulega forrettindabundin. Ekki eingongu endurheimtur. Verkfaeri sem neer 90% endurheimtum a 22,7% naekvaemi tekur meiri tiltaeka efni. En thad skapar yfirfarnar byrdi fyrir 77,3% ranga flagga. Naer ssu yfirfarning gerist ekki, fylgir breidt of-geymsla.

Hver merking i framleidningu er krafa til domstolsins. Hann seggur: thetta efni er legitima geymt. Eftir Schnitzer Steel verdur ssu krafa ad standast.

Fyrir naenar upplysingar um hvernig nafnleynd-verkfaeri eru olikur hefdbundnum PII-greiningu, sja leidsogn um gervigreind naekvaemi i lagaskjalayfirfarning. Til samhengi um forrettindaskrar og gervigreind-verkfaeri, sja grein um loerfaeda-bidhafi forrettindi og gervigreind.

Heimildir

Ertu tilbúinn að vernda gögnin þín?

Byrjaðu að anonymiza PII með 285+ gerðum í 48 tungumálum.

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.