ScreenJournal

United States — California and federal law

Version 1.0 · Effective 28 October 2026

This page explains how ScreenJournal, operated by Cyberinfra Limited (Isle of Man) ("we", "us"), and the employers who use it handle personal data and monitoring under United States law, with California as the main case because it has the most demanding rules. It is written for our customers, for the people whose work activity they monitor, and for people who take part in calls with them.

This is a plain-language explainer and is not legal advice. Our Privacy Policy describes in full what we collect and why; this page adds what is specific to US and California law. Laws named here were checked in October 2026; several take effect during 2027, and we say so where that is the case.

1. Which laws apply

The United States has no single workplace-monitoring statute. An employer that deploys ScreenJournal in the US is governed by a layered set of rules:

  • Privacy law. The California Consumer Privacy Act ("CCPA") applies to employee data where the employer is a "business" (section 4). The other nineteen state privacy laws in force in 2026 exempt employee and business-contact data, so for monitoring data they do not apply.
  • Recording law. California's Invasion of Privacy Act and the recording laws of about a dozen other states require the consent of every party before a conversation is recorded (section 8). Federal law requires one party's consent.
  • Workplace privacy and surveillance law. California's constitution, Labor Code and three new statutes taking effect in 2027 (section 7), and the monitoring-notice statutes of New York, Connecticut, Delaware and Maine (section 13).
  • Employment law. Federal and California wage-and-hour law, which decides what time must be paid (section 10); the National Labor Relations Act (section 11); and anti-discrimination law, including California's rules on automated-decision systems (sections 6 and 12).
  • Sector rules for health, payment-card, financial and education data, and the Fair Credit Reporting Act (section 14).

Employers with staff in several states usually apply the strictest standard to everyone. In the US that is California.

2. Who plays which role

  • Your employer decides. It chooses whether to use ScreenJournal, which features to turn on, what to monitor and why, and what to do with the results. Under the CCPA it is the "business" that collects your data; under the Labor Code and the Fair Employment and Housing Act it is your employer. It is your first point of contact.
  • Cyberinfra Limited is a service provider. We process monitoring data on the employer's behalf, under a written contract and on its instructions, in the sense of section 1798.140(ag) of the California Civil Code. We do not sell or share personal information, we do not use monitoring data for our own purposes, and we do not use it to train AI models. Our Data Processing Agreement contains the US state privacy law annex (Annex 4) under which we act as a service provider.
  • We are the "business" only for the data we control directly, such as customer account, billing, sign-in security and website data. As at the date of this page we do not meet any of the thresholds that make a company a "business" under the CCPA; we apply the rights in section 4 to that data anyway, as our Privacy Policy describes.

Data protection contact. Our Data Protection Officer is a role, not yet a named individual. Until an officer is appointed, the role is reached through support@screenjournal.ai with the subject line "Data protection request". See our Data Protection Contacts page.

3. What the product does

The ScreenJournal desktop app runs on Windows and macOS and is visible in the menu bar or system tray while it runs. While tracking is on it records each connected display as short video segments, by default at one frame per second. The app paints the cursor position and clicks into the recording as markers. Applications and sites on the organisation's exclusion list are not captured. The app also records application names, window titles, browser addresses, and presence and idle signals.

  • Default mode. Screen video is uploaded for analysis and deleted after analysis; any temporary copy is removed by a storage lifecycle rule we configure on the bucket. The desktop app keeps its own copies on the member's device for the period the organisation sets, up to three months.
  • Record + Save (employer-elected). If the organisation enables it, screen video is stored on our servers so managers can play it back.
  • Alert evidence clips (employer-elected). Where the organisation's alert rules call for it, a short clip is kept as evidence for a flagged event.
  • Audio is off by default (see section 8).
  • Timesheet and pay. The Timesheet page computes pay figures from the employer's pay rates. It counts only time the service scored neutral or productive, not time scored unproductive or marked idle, and lets a manager exclude further days or segments (see section 10).
  • Connected AI tools. If the organisation connects its own AI tool through our integration feature, that tool can read the organisation's data within the connecting user's permissions.
  • Sign-in region. At desktop sign-in we derive a country or region from the network address, using a local database, and store it on the user record.

At every desktop sign-in the app shows a notice that it may record the screen and, where the organisation turns it on, meeting audio. Your acknowledgment is recorded with the notice's version and the time; if the record fails, sign-in still completes. Monitoring does not wait for it.

4. The California Consumer Privacy Act

When it applies. The CCPA applies to a for-profit company doing business in California that, in the previous year, had gross revenue above USD 26,625,000 (the figure in force since 1 January 2025), or bought, sold or shared the personal information of 100,000 or more consumers or households, or earned half or more of its revenue from selling or sharing personal information. An employer below all three thresholds is not a "business" and the CCPA does not apply to its monitoring. The recording, workplace and wage rules in the rest of this page apply to every employer whatever its size.

Employee data is covered. The CCPA's partial exemptions for employee, applicant, contractor and business-contact data expired on 1 January 2023. A California employer that is a business owes its workers the full set of CCPA duties for monitoring data, and the California Privacy Protection Agency (the "CPPA") has enforced them: its September 2025 order against Tractor Supply concerned, among other things, a deficient notice to job applicants.

Notice at collection. Before or when it starts collecting, the employer must tell each worker the categories of personal information collected, including any sensitive categories, the purposes, whether the data is sold or shared, and how long each category is kept. Under the CPPA's regulations, no notice means no collection. The notice is the employer's; our Workplace Monitoring Notice and Privacy Policy list every category, purpose and retention period and can be used as inputs, but they do not replace it.

Sensitive personal information. Screen content can incidentally include categories the CCPA treats as sensitive: the contents of a person's email and messages where the employer is not the intended recipient, health information, account credentials, and signs of union membership or religious belief. The CCPA gives a person the right to limit the use of sensitive information where it is used to infer characteristics about them, which productivity scoring arguably is. The exclusion list and the sensitive-category packs (banking, payroll, HR, health and adult content, on by default) narrow what is captured; an employer's notice should still name these categories where they may be collected.

Your rights. Where the CCPA applies, you have the right to know what personal information has been collected about you, including inferences such as scores and flags; to delete it; to correct inaccurate information; to limit the use of sensitive personal information; and not to be retaliated against for exercising these rights. A business must respond within 45 days, which it may extend once. California does not give a right of appeal to the business; several other states do.

How requests are fulfilled today. Your employer is the business, so raise requests about monitoring data with your employer first. If you write to us, the CPPA's regulations require us to act on your employer's instructions or to tell you that the request must go to the employer; we pass it on and help the employer respond. We honour access, correction and deletion instructions manually; there is no self-service export. Where your employer has enabled it, you can redact a time range from your own timeline; a redaction removes the activity data for that range, including any Record + Save video. Alert evidence clips are not removed by a member's redaction request.

What we do as a service provider. The CCPA and its regulations require the contract between a business and its service provider to contain specific terms: a limited and specified purpose, no selling or sharing, no use outside the direct business relationship, no combining with other data, the same level of privacy protection the business owes, assistance with consumer requests, audits and risk assessments, notice if we can no longer comply, and the right to stop and remediate unauthorised use. Our Data Processing Agreement contains service-provider terms (Annex 4) and the assistance, audit, subprocessor and deletion provisions these rules call for. We do not use monitoring data to build or improve services for anyone other than the employer whose data it is.

Penalties. The CPPA and the Attorney General may impose up to USD 2,663 per violation, or USD 7,988 per intentional violation (figures in force since 1 January 2025), counted per affected person; there is no longer an automatic cure period.

5. Risk assessments and automated decision-making under the CPPA's 2026 rules

The CPPA's regulations on risk assessments, cybersecurity audits and automated decisionmaking technology took effect on 1 January 2026. They bind businesses; we describe below what they ask of an employer that is a business and what we do to help.

Risk assessments. A business must complete a documented risk assessment before it uses automated processing to infer a worker's performance, reliability or behaviour from "systematic observation" of that worker, a term the regulations define to include continuous video or audio recording. Continuous screen recording with AI scoring falls within that description. Processing sensitive personal information for monitoring, and using automated decision-making technology for a significant decision, are further triggers. Processing that began before 2026 must be assessed by 31 December 2027; an executive attestation and summary go to the CPPA by 1 April 2028 and yearly after that. Service providers must cooperate and supply the facts the business needs. Our Automated Processing Statement, Security and Subprocessors pages are written to be those inputs: what is inferred, from which data, by which model, where it goes wrong, and who processes it. We will answer a customer's written questions for its assessment.

Cybersecurity audits. Annual independent audits apply only to businesses that meet the CCPA revenue test and process the data of 250,000 or more consumers (50,000 for sensitive data), or that earn half their revenue from selling personal information, with the first reports due from 1 April 2028. Most of our customers will not be in scope.

Automated decisionmaking technology ("ADMT"). The regulations treat technology as ADMT where its output is used to make a decision "without human involvement", and they list, as significant decisions, the allocation or assignment of work, hourly or per-assignment pay and bonuses, promotion, demotion, suspension and termination. The CPPA's own example of a consolidated notice is "an employer's proposed use of productivity monitoring software to determine the employee's allocation/assignment of work and compensation, and to determine which employees will be demoted." Human involvement means a reviewer who knows how to read the output, reviews it with other relevant information, and has the authority to change the decision. A business that uses ADMT for a significant decision must, from 1 January 2027, give a pre-use notice, offer a way to opt out (or, for discipline and dismissal, a human appeal to someone who can overturn the decision), and answer requests for an explanation of the logic and the outcome.

The service draws five kinds of automated inference about the people it monitors:

  1. an activity score from 1 to 5 for each recorded segment of screen activity;
  2. a productivity percentage, measured against the working day;
  3. a position in a weekly ranking of the organisation's members;
  4. flags that activity may be simulated, for example by a mouse-jiggling device;
  5. matches against alert rules written by the organisation.

Segment scores and descriptions are produced by a Google Gemini model accessed through Google Vertex AI, and alert evaluations by a Google Gemini model accessed through Google's Gemini API, from screen content and activity signals such as presence and idle time. The productivity percentage and the weekly ranking are calculated from those scores. Simulated-activity flags come from the same model reviewing the segment's screen recording, on which the desktop app marks the cursor position and clicks; fixed checks in our code then discount the device's own idle periods and apply minimum thresholds before a flag is set. These outputs are probabilistic. They can be wrong, and they describe what was on screen, not the quality or value of anyone's work.

ScreenJournal makes no employment decision. Any decision about a person, whether on pay, performance, discipline or anything else, is made by your employer's managers, who are responsible for it.

The following human-review surfaces exist today:

  • the Review page, where managers can examine flagged segments, and the alerts inbox, where a manager approves or rejects a member's explanation;
  • corrections, where the organisation's administrators change a segment's score;
  • Add a reason, where a member, or a manager on the member's timeline, explains a paused, offline or away stretch (when the organisation's policy allows manual entries);
  • the member's own Activity timeline, which shows their segments, descriptions and scores.

If you disagree with a score, flag or ranking, ask your employer first. Your employer can correct the record or redact a time range. If your employer does not respond, contact support@screenjournal.ai.

A flag that activity may be simulated is an indicator that calls for human judgement. It is never proof of misconduct, and it should not be relied on without a person reviewing the underlying activity.

How this maps onto ScreenJournal. Whether an employer is "using ADMT" depends on its own practice, not on the tool. A manager who reviews a score, a ranking or a flag together with the underlying activity, and who can change the outcome, supplies the human involvement the rules describe; applying an output to pay or discipline without that review does not. Bulk actions, such as excluding every low-scored segment from pay in one step, deserve particular care. If an employer does rely on outputs without human involvement, it must meet the ADMT rules itself; we help by acting on its instruction to stop monitoring a worker who opts out, and by supplying the data and explanations it needs to answer an access request.

6. Discrimination law and automated-decision systems

California's Civil Rights Council regulations on automated-decision systems in employment took effect on 1 October 2025. They apply to every employer with five or more employees, whatever its revenue. An "automated-decision system" is any computational process that makes a decision, or facilitates a human decision, about an employment benefit, and pay is an employment benefit. ScreenJournal's scores, rankings, flags and pay figures are such a system, however they are used.

  • The employer answers for discriminatory outcomes. If reliance on scores, rankings or flags results in discrimination on a protected basis, including through a proxy closely correlated with one, the employer is liable in the ordinary way. Evidence of anti-bias testing, or its absence, is relevant to a claim or defence. Idle-time and activity-rate metrics may correlate with disability, pregnancy or caregiving; reasonable accommodation is the employer's duty, and any medical information that appears on screen must not be used against a person.
  • Records must be kept for four years. Data used in or produced by the system must be kept for four years from the record or the personnel action. We keep monitoring data for the subscription term and delete it within sixty days after it ends; an employer that must keep longer should export reports and timesheets before then (see our Retention & Deletion Protocol).
  • Vendors as agents. A third party that itself carries out a regulated activity on the employer's behalf, such as deciding pay through an automated system, can be treated as the employer's agent. ScreenJournal makes no employment decision: the employer configures the service, reviews its outputs and decides. Our Terms and Acceptable Use Policy require human decisions, forbid alert rules that target protected characteristics, and forbid decisions based solely on automated output.

Federal law applies alongside: Title VII (disparate impact), the Americans with Disabilities Act (no disability-related inquiries or screening standards that are not job-related; accommodation), the Age Discrimination in Employment Act and the Pregnant Workers Fairness Act. The Equal Employment Opportunity Commission withdrew its AI guidance documents in 2025; the statutes are unchanged and private claims remain available.

7. New California workplace laws taking effect in 2027

Three statutes signed on 30 September 2026 apply directly to monitoring software. As we understand them from published summaries, pending the official texts:

  • SB 947, the No Robo Bosses Act of 2026 (Labor Code sections 1520 to 1526.1, operative 1 July 2027). An employer may not rely solely on an automated-decision system to discipline or dismiss a worker. Where it relies primarily on such a system, a person must corroborate the output against the underlying information; output that cannot be corroborated, or that the person finds inaccurate, incomplete or misleading, must not be acted on. The worker gets a plain-language written notice that a system was primarily relied on and a human contact, and may request a description of the data used. Systems may not be used to infer protected status or to penalise the exercise of legal rights. Civil penalty USD 500 per violation. ScreenJournal's Review page, corrections and explanations are built for this kind of human check, but the corroboration and the notice are the employer's.
  • AB 1883 (Labor Code sections 1580 to 1582, from 1 January 2027). Employers may not use AI workplace-surveillance tools to recognise, infer or predict a worker's emotional state, or to collect neural data, for any purpose. ScreenJournal collects no neural data, and the service is not designed to infer anyone's emotional state: the model scores observable on-screen activity, not mood. Employers must not write alert rules that ask for it, and our Acceptable Use Policy already forbids inferring health, political or private-life matters.
  • AB 1331 (Labor Code section 1560 and following, from 1 January 2027). No "workplace surveillance tool", a term broad enough to include ScreenJournal, may monitor workers in a workplace bathroom, and workers may leave such tools behind before entering one. The desktop app records only while signed in and tracking is on, and a timed pause is always available.

The Governor's signing statements asked the Legislature to clarify definitions in AB 1883 and AB 1331, so amendments in 2027 are likely; we will update this page as they are enacted.

8. Audio and California's recording law

California requires the consent of every party to a recorded conversation. Under Penal Code section 632 it is a crime, and a civil wrong, to record a confidential communication without the consent of all parties; a communication is confidential where any party reasonably expects it is not being recorded. Under section 632.7 the same applies to any communication, confidential or not, where any participant is on a cellular or cordless phone, and the California Supreme Court held in Smith v. LoanMe (2021) that this binds the parties to the call, not just eavesdroppers. The civil remedy under section 637.2 is USD 5,000 per violation or three times the actual damage, whichever is greater, with no need to prove harm; the criminal penalty is up to USD 2,500 and a year in county jail per violation, more for repeat offenders. Evidence obtained in breach is inadmissible in any proceeding. In Kearney v. Salomon Smith Barney (2006) the Supreme Court held that California law applies to the recording of a Californian by a business elsewhere, so what matters is where the people on the call are, not where the employer or we are based. The federal Wiretap Act requires only one party's consent and is a floor, not a ceiling.

Three points from the case law decide how the audio feature can be used in California:

  • The monitored employee's consent is not the other party's consent. A company that records its own staff's calls with customers still needs the customer's consent (Kight v. CashCall, 2011).
  • Recording only the employee's side of a VoIP call is still recording (Gruber v. Yelp, 2020).
  • A recording tool that keeps or reuses conversations for its own purposes can be liable as a third-party eavesdropper, not merely as its customer's tool (In re Otter.AI, N.D. Cal., August 2026). We do not keep or use recordings for any purpose of our own, and our model providers are bound not to train on them.

How audio works in ScreenJournal:

  • Audio is off by default. An owner or administrator must turn it on for the organisation, accepts the Call-Recording Addendum when doing so, and is shown a warning about recording laws.
  • There is no per-call announcement. Nothing in the product tells the other people on a call that they are being recorded or stops capture until they agree.
  • When audio is on, the app records whenever a listed meeting or calling application uses the microphone. It records all participants, including customers and other people outside the organisation, and where listen capture is enabled it records whole meetings and webinars.
  • Transcripts attribute each passage to the "member" or to the "other party". The transcription model also separates the voices it hears and may label a passage with a name spoken in the conversation; that label is stored with the transcript, and the Audio page shows only the member's name or "Other party". We do not match voices against voiceprints or identify anyone biometrically.
  • The member sees "Recording from" and the application's name while audio is recording.

Telling every participant and obtaining their consent is the responsibility of the organisation that turned audio on. The method courts accept is a clear statement at the start of the call, before recording begins, with the participant free to object or leave; for calls over the public telephone network, the Public Utilities Commission's General Order 107-B also recognises a periodic beep tone audible to all parties. A line in a meeting invitation is a supplement, not a substitute. Listen capture of webinars should stay off unless the organiser's terms and the law allow every attendee to be recorded. We recommend that organisations with members or participants in California keep audio off until they have confirmed their method with counsel.

The same all-party rule applies in Connecticut (civil), Delaware, Florida, Illinois, Maryland, Massachusetts, Montana, Nevada, New Hampshire, Oregon, Pennsylvania and Washington, each with its own penalties. Elsewhere in the US, one party's consent suffices, but the employer still owes the notices in section 13.

Silent screen capture of a video call shows the other participants' names and pictures without sound. California's courts are divided on whether soundless video is a "communication" under section 632; the people shown can read our notice for people recorded on calls and webinars.

9. Screen recording, privacy and the Labor Code

Article I, section 1 of the California Constitution gives every person, including an employee of a private company, a right of privacy. A claim requires a reasonable expectation of privacy and a serious invasion; advance notice and the chance to consent reduce that expectation, and the courts have upheld workplace video surveillance that is narrowly tailored in place, time and scope and prompted by a legitimate business concern (Hernandez v. Hillsides, 2009), and found no privacy in a company computer used at home where the employee had acknowledged a monitoring policy. The notice the employer gives, and the acknowledgment the app records at sign-in, matter for this reason. What the employer records should be proportionate to its purpose; the exclusion list, leaving Record + Save and audio off, and limiting monitoring to working hours all narrow it.

Specific rules apply:

  • Labor Code section 435 forbids audio or video recording of employees in restrooms, locker rooms and changing rooms, and Penal Code section 647(j) makes concealed recording in private places a misdemeanour. The desktop app records the screen, not the room, and only while signed in and tracking is on.
  • Labor Code section 980 forbids an employer from requiring access to a worker's personal social media. Personal sites belong on the exclusion list, and content captured from a personal session must not be used against the worker.
  • Labor Code section 2802 requires an employer to reimburse the reasonable cost of personal devices and home internet it requires a worker to use. An employer that asks staff to run the app on their own computers owes that reimbursement.
  • Labor Code section 1198.5 lets current and former employees inspect records relating to their performance within 30 days of asking. Stored scores, rankings and flags are likely to be such records; the employer can meet a request from the member's own timeline, the reports and the timesheet export. Section 432 entitles a worker to a copy of any document they signed, which includes the sign-in acknowledgment.
  • Labor Code sections 1102.5 and 98.6 protect a worker who objects to unlawful monitoring or complains about unpaid wages; monitoring must not be intensified or used against them for it.

10. Pay, hours worked and the Timesheet

This section matters most. Federal and California law require an employer to pay a non-exempt (hourly) employee for all hours worked: all time the employee is suffered or permitted to work, and, in California, all time the employee is under the employer's control, whether or not productive. Waiting, thinking, reading, short rest breaks of up to twenty minutes and a few idle minutes while on duty are all working time (29 C.F.R. part 785; Industrial Welfare Commission wage orders, section 2(K)). California does not excuse regularly recurring minutes as trivial (Troester v. Starbucks, 2018), and an employer that holds a frame-by-frame record of the day cannot say it did not know the work was done. Only time the employee was fully relieved of duty, such as a bona fide meal period of thirty minutes or more, is unpaid.

ScreenJournal's activity scores, productivity percentage, idle signals and simulated-activity flags describe what was on screen. They do not establish that an employee was relieved of duty, and they must not be used to reduce a non-exempt employee's pay for time worked. In particular:

  • the Timesheet's hours count only time scored neutral or productive and not idle, so for a non-exempt employee they are not hours worked; the time scored unproductive or idle during the working day must be added back before the figures are used for that person's pay;
  • the Timesheet's segment and day exclusions, including "Exclude flagged", must not be applied to a non-exempt employee's working time; low-scored, idle-marked and flagged stretches of the working day are still hours worked;
  • the three-minute idle threshold is a presence signal, not a pay rule; where it marks time away, the employee's "reason" and the employer's own knowledge decide whether the time was worked;
  • wages already paid may not be clawed back for low scores (Labor Code sections 221 and 224);
  • docking an exempt (salaried) employee's pay for scores or idle time is an improper deduction that can cost the employer the exemption for the whole class;
  • the employer, not ScreenJournal, must keep accurate records of hours worked and issue wage statements showing them (Labor Code sections 226 and 1174; 29 C.F.R. part 516). The Timesheet is computed from scored activity and the employer's settings; it is not a payroll record; and
  • monitoring independent contractors with a locked recorder, idle thresholds and rankings is evidence of control under California's ABC test, which presumes a worker is an employee.

Exposure for getting this wrong falls on the employer: unpaid wages and overtime, liquidated damages, waiting-time and wage-statement penalties, Private Attorneys General Act penalties and attorney's fees, in class or collective actions. The product gives the employer the gross tracked record; every exclusion is the employer's own, logged act.

11. Unions and the National Labor Relations Act

Monitoring that targets, or gives the impression of targeting, union or other protected concerted activity violates section 8(a)(1) of the Act, in any state. In a unionised workplace the introduction and use of monitoring technology is a mandatory subject of bargaining (Colgate-Palmolive, 1997). Alert rules must not be written to watch organising, pay discussions or complaints; our Acceptable Use Policy forbids it. The General Counsel memorandum of 2022 that proposed a presumption against intrusive monitoring was rescinded in February 2025; the underlying Board law described here still applies.

12. Your rights as a monitored worker in the United States

  • Notice. Your employer must tell you about monitoring before it starts where state law requires it (California, where the CCPA applies; New York, Connecticut, Delaware and Maine everywhere). The sign-in notice confirms you were shown our notice; it is not your employer's.
  • Access, deletion and correction under the CCPA where your employer is a business, and access to performance records under Labor Code section 1198.5 in any case.
  • Human review of any decision taken with automated outputs, and from 1 July 2027 the corroboration and notice rights in SB 947.
  • Pay for all hours worked, whatever your scores.
  • No retaliation for exercising any of these rights.

Raise requests with your employer first. If you write to us, email support@screenjournal.ai with the subject line "Data protection request"; we verify who you are, pass the request to your employer and help it respond. You can complain to the California Privacy Protection Agency or the Attorney General (privacy), the Labor Commissioner (wages, notice and surveillance statutes), the Civil Rights Department (discrimination) and the National Labor Relations Board (protected activity), or the equivalent bodies in your state.

13. Other states

Monitoring-notice statutes. Four states require employers to notify workers before electronic monitoring, as we understand them:

StateLawWhat the employer must doPenalty
New YorkCivil Rights Law section 52-c (since 7 May 2022)Prior written notice on hiring that calls, email and internet use may be monitored "at any and all times and by any lawful means", acknowledged in writing or electronically, plus a conspicuous postingUSD 500, 1,000, then 3,000 per offence, enforced by the Attorney General
ConnecticutGeneral Statutes section 31-48d, amended by Public Act 26-73 from 1 October 2026Prior written notice of the types of monitoring and, from October 2026, the specific locations where it occurs, posted conspicuously; a plain-language statement to new hiresUSD 500, 1,000, then 3,000, by the Labor Commissioner
DelawareTitle 19, section 705Either an electronic notice each day the worker uses the employer's email or internet, or a one-time written or electronic notice acknowledged by the workerUSD 100 per violation
MainePublic Law 2025, chapter 524 (LD 61), in force from July 2026Notice before monitoring starts, disclosure to applicants and a yearly written notice; no surveillance in a worker's home, personal vehicle or personal property unless the job requires it; workers cannot be required to install monitoring software on personal devicesUSD 100 to 500 per violation, by the Department of Labor

New York and Delaware accept an electronic acknowledgment, so the record the app keeps at desktop sign-in can evidence the acknowledgment if the employer's own notice says what the statute requires. It cannot satisfy the posting New York and Connecticut require, Maine's annual and applicant notices, or Connecticut's location disclosure. Employers should treat the sign-in acknowledgment as a tool for their own notice, not a discharge of it.

Automated decisions. Illinois makes it a civil-rights violation, from 1 January 2026, to use AI in employment decisions with a discriminatory effect, to use zip codes as a proxy, or to fail to notify workers that AI is used. Colorado's AI Act was repealed and replaced in May 2026 by a narrower notice-and-explanation law (SB 26-189) in force from 1 January 2027; it requires deployers to tell people when automated decision-making technology materially influences an employment decision and to explain adverse decisions, and requires developers to document their systems for deployers; our Automated Processing Statement is the documentation we provide for that purpose. New York City's bias-audit law covers hiring and promotion tools only.

Biometrics. Illinois, Texas, Washington and Colorado regulate voiceprints and other biometric identifiers. ScreenJournal creates no voiceprint and identifies no one biometrically; a model-assigned speaker label in a transcript is a label, not a template. Employers must not add biometric identification through alert rules or connected tools.

14. Regulated data: health, payment cards, finance, education and background checks

  • Health data (HIPAA). A vendor that stores protected health information on a covered entity's behalf is a business associate whether or not it signs an agreement, and a covered entity that uses such a vendor without one is itself in breach. ScreenJournal is not designed to process protected health information and we do not offer a business associate agreement. The health pack is on the exclusion list by default; a healthcare customer must keep it on, add every system that shows patient information, and keep Record + Save and audio off for clinical work. If we discover protected health information in our systems we will delete it and tell the customer.
  • Payment cards (PCI DSS). Card numbers and security codes must never be recorded; our Terms require payment applications to stay on the exclusion list and recorded calls to use DTMF masking. Before analysis output is stored, an automated filter removes card numbers from descriptions and window titles.
  • Financial and education data. Banks, lenders, insurers and brokers (GLBA), schools (FERPA) and substance-use treatment programmes (42 C.F.R. part 2) must bind or exclude their vendors; we are not a "school official" or a qualified service organisation, so those customers must exclude the relevant systems.
  • Background checks (FCRA). ScreenJournal is not a consumer reporting agency and its outputs are not consumer reports. Every input comes from the employer's own devices and the employer's own experience with its own staff; we add no third-party data and do not train shared models on one organisation's workers to serve another. Employers must not use outputs as a background check or combine them with third-party data for hiring or eligibility decisions. A federal circular that treated some AI productivity scores as consumer reports was withdrawn in May 2025; private litigation on the point continues, and we will update this page if the law moves.

15. Where your data goes

Since 20 October 2026 the service, including its database, is hosted in India, in Amazon Web Services' Mumbai region, where activity time-series data is also stored; before that date the servers were in Germany. Some data is transferred onward to the vendors listed on our Subprocessors page, including AI processing through Google Vertex AI and Google's Gemini API, and US vendors for the web app, email and crash reports. US law imposes no general restriction on sending personal data abroad; an employer's notice at collection should say where its workers' data goes. Cyberinfra Limited is established in the Isle of Man, whose data protection law applies the GDPR to everything we process wherever our customers are, and the Isle of Man Information Commissioner is our home regulator.

16. Security breaches

California's breach law, amended from 1 January 2026, requires a business to notify affected residents within 30 calendar days of discovering a breach and to send a sample notice to the Attorney General within 15 days where more than 500 residents are affected. A company that holds data it does not own, which is our position for monitoring data, must notify the owner immediately following discovery. Under our Data Processing Agreement we notify the affected customer without undue delay after becoming aware of a breach, and in any event within 72 hours of confirming it, and give it the facts it needs for its own notices. We do this manually; there is no automated notification. A person whose unencrypted data is exposed because a business failed to maintain reasonable security can sue the business for USD 107 to 799 per consumer per incident. Our Security page describes the measures in place and what is not yet built.

17. Our own duties to Californians

For visitors to our website and the people who administer or pay for an organisation, we are the company that collects the data. Our Privacy Policy is written to meet the California Online Privacy Protection Act, and our Cookie Policy explains that analytics and advertising tags load only after consent. We do not sell personal information and we do not disclose it to third parties for their direct marketing. We do not yet read Global Privacy Control or Do Not Track browser signals automatically; use the cookie banner to withhold or withdraw consent. The support chat on our website is an AI assistant and is labelled as one in the chat window. A person who buys a subscription for personal use is a consumer under California's automatic-renewal law and can cancel online from the Billing page at any time.

18. Retention

Data is retained for the term of your employer's subscription unless it is deleted earlier on a verified request or by your employer. Where Record + Save is enabled, stored screen video is kept for up to three months by policy. Automated expiry is not yet built; deletion is carried out by our operational procedure.

Our Retention & Deletion Protocol sets out how long each kind of data is kept.

We list every change on the Legal page with its effective date and, where practicable, post material changes before they take effect. Several of the laws on this page take effect in 2027; we will update it as they do.

Related: Privacy Policy, Terms of Use, Data Processing Agreement, Call-Recording Addendum, Workplace Monitoring Notice, Your Rights as a Monitored Employee, Automated Processing Statement, Subprocessors, Security, Retention & Deletion Protocol, Data Protection Contacts and Cookie Policy.

Contact: support@screenjournal.ai (subject "Data protection request"). Post: Cyberinfra Limited, 50 Athol Street, Douglas, Isle of Man IM1 1JB.

This page is for information and is not legal advice. Employers remain responsible for the lawfulness of their own monitoring, for every notice and consent the law requires, for recording law on any recorded call, and for paying their staff for all hours worked.

Changes and previous versions

  • 28 October 2026v1.0: first publication — CCPA employee data and service-provider role; the CPPA's 2026 risk-assessment and automated-decision rules; FEHA automated-decision regulations; SB 947, AB 1883 and AB 1331 from 2027; California recording law and the audio feature; constitutional and Labor Code privacy rules; pay for all hours worked; NLRA; other states' notice, consent, biometric and AI laws; regulated data; breach notification.

Questions about this document: support@screenjournal.ai. Canonical URL: /legal/united-states-california.