Candidates do cheat on online skills assessments, and generative AI made the most common formats easier to outsource, particularly coding challenges and knowledge tests. A 2024 study found ChatGPT-3.5 solved 71.43 percent of 98 archived LeetCode problems, and all three models tested earned HackerRank problem-solving certificates at basic and intermediate levels [7]. Proctoring narrows the gap without closing it, since the meta-analysis most cited in this area, covering 49 studies and 100,434 test takers, put unproctored scores ahead by a standardized mean difference of only 0.20 [8], and vendors say plainly that a candidate reading from a second monitor or phone is the case their software cannot see [1]. Detection accuracy is vendor-published, including HackerRank's stated 85 percent precision on its AI plagiarism model, which means roughly one flagged session in seven is wrong [3]. Assessment design carries better evidence than surveillance, since structured interviews ranked as the top selection procedure in the 2022 re-analysis of personnel selection validity [10].
\n","summary":"Cheating on remote skills assessments is real, and generative AI widened it for coding challenges and knowledge questions. Proctoring catches tab switching, pasted code, proxy test takers, and leaked items, and it cannot see a phone outside the webcam frame. Every published detection rate is a vendor figure. Biometric identity checks and recorded video also carry Illinois BIPA and AI Video Interview Act exposure.","publishedAt":"2026-07-21T21:20:42.255","verifiedAt":"2026-07-21T00:00:00","editorialStatus":"APPROVED","lastReviewedAt":"2026-07-21T00:00:00","nextReviewDueAt":"2026-10-21T00:00:00","templateVersion":"v2","aliases":["Do candidates cheat on online assessments?","Does proctoring stop cheating on skills tests?","Can assessment software detect ChatGPT?","How do candidates cheat on remote coding tests?","Is online proctoring effective for hiring?","Anti-cheat features in pre-employment assessments","How common is cheating on pre-employment tests?","Can HackerRank detect AI-generated code?","Proctoring vs assessment design for test integrity","Are remote skills assessments still valid with AI?","Legal risks of proctoring job candidates","How to stop AI cheating in technical screening"],"confidenceScore":78,"confidenceLabel":"Medium","canonicalUrl":null},"contributor":{"id":"ec39deab-44fe-48d8-9029-fefe993ab85a","slug":"answer-stack","displayName":"AnswerStack","websiteUrl":null},"contributorOrganizationProfile":{"entityId":"ec39deab-44fe-48d8-9029-fefe993ab85a","legalName":null,"description":null,"websiteUrl":null,"imageUrl":null,"slogan":null,"subtitle":null,"facts":[],"coiNote":null,"foundingDate":null,"numberOfEmployeesText":null,"contactPoint":null,"address":null,"headquartersText":null,"organizationType":null},"contributorPerson":{"slug":"answerstack-editorial-team","displayName":"AnswerStack Editorial Team"},"sections":[{"id":"da638736-553a-42f2-bc79-c713e6eefaa2","sectionKey":"how_common_is_it","sectionType":"markdown_section","heading":"How common is cheating on online skills assessments?","introMarkdown":"Cheating on unsupervised online assessments happens often enough that every major vendor sells detection features, though honest prevalence numbers are harder to pin down than the marketing suggests. CodeSignal announced in February 2026 that flagged fraud attempts across its platform rose from 16 percent of assessments in 2024 to 35 percent in 2025, reaching 40 percent at entry level [6]. HackerRank has separately stated that 25 percent of technical assessments show signs of plagiarism [4]. Both come from companies selling anti-cheat features, neither publishes a sample size or a method, and both count flags rather than confirmed cheating. Other widely repeated survey statistics were left out because no openable primary source stood behind them.\n\n### What generative AI changed\n\nA 2024 study tested ChatGPT-3.5, Gemini 1.0 Pro, and Meta AI against archived programming problems. ChatGPT solved 71.43 percent of 98 archived LeetCode problems, Gemini 68.37 percent, and Meta AI 58.16 percent, and all three obtained HackerRank problem-solving certificates at basic and intermediate levels [7]. Difficulty mattered, since the same models cleared only 8 of 49 problems across nine live contests [7]. Those models are two generations old now, so current systems clear the easier tiers more comfortably.\n\n### What proctoring changes in measured scores\n\nProctoring moves scores less than most buyers expect. The meta-analysis of 49 studies and 100,434 test takers cited in a 2025 review found unproctored scores higher by a standardized mean difference of 0.20 [8]. The accompanying experiment with 226 participants found no significant main effect of proctoring condition and observable cheating in only 3 of 74 recorded participants [8]. Both settings were low stakes, and a job offer raises the incentive considerably.","introHtml":"Cheating on unsupervised online assessments happens often enough that every major vendor sells detection features, though honest prevalence numbers are harder to pin down than the marketing suggests. CodeSignal announced in February 2026 that flagged fraud attempts across its platform rose from 16 percent of assessments in 2024 to 35 percent in 2025, reaching 40 percent at entry level [6]. HackerRank has separately stated that 25 percent of technical assessments show signs of plagiarism [4]. Both come from companies selling anti-cheat features, neither publishes a sample size or a method, and both count flags rather than confirmed cheating. Other widely repeated survey statistics were left out because no openable primary source stood behind them.
\nA 2024 study tested ChatGPT-3.5, Gemini 1.0 Pro, and Meta AI against archived programming problems. ChatGPT solved 71.43 percent of 98 archived LeetCode problems, Gemini 68.37 percent, and Meta AI 58.16 percent, and all three obtained HackerRank problem-solving certificates at basic and intermediate levels [7]. Difficulty mattered, since the same models cleared only 8 of 49 problems across nine live contests [7]. Those models are two generations old now, so current systems clear the easier tiers more comfortably.
\nProctoring moves scores less than most buyers expect. The meta-analysis of 49 studies and 100,434 test takers cited in a 2025 review found unproctored scores higher by a standardized mean difference of 0.20 [8]. The accompanying experiment with 226 participants found no significant main effect of proctoring condition and observable cheating in only 3 of 74 recorded participants [8]. Both settings were low stakes, and a job offer raises the incentive considerably.
\n","outroMarkdown":null,"outroHtml":null,"contentJson":{},"configJson":{},"noteMarkdown":null,"noteHtml":null,"sortOrder":10},{"id":"b4bcaff8-8626-4d0c-b460-07199ea0a0ab","sectionKey":"how_candidates_cheat","sectionType":"markdown_section","heading":"How do candidates actually cheat on a remote assessment?","introMarkdown":"Candidates use four broad methods, and only some of them leave any trace inside the assessment software.\n\n### A second physical device running a chatbot\n\nA phone or second laptop beside the test window is the hardest method to detect, because nothing on that device touches the assessment session. TestGorilla, describing the limits of its own detection, notes that candidates can read from secondary devices or monitors the software does not track [1]. Webcam proctoring sees a face and a rough gaze direction, not what sits two feet to the left.\n\n### An invisible overlay on the test machine\n\nA second class of tool renders answers in a layer that never appears in a shared screen or recorded video feed. Cluely launched in 2025 around a hidden in-browser window the company called undetectable, prompting two startups to ship counter-detection products within weeks, and its chief executive answered by floating smart glasses or a recording necklace as ways past detection [11].\n\n### Someone else takes the test\n\nProxy test-taking and identity fraud are two of the four violation categories CodeSignal says its systems target [5]. The variant that defeats a webcam is a coached candidate, where the right person sits on camera while a helper watches a shared screen and feeds answers by phone.\n\n### Circulated answers from leaked question banks\n\nCodeSignal runs a service called LeakSweep that monitors sites including Stack Overflow and LeetCode for leaked questions, and HackerRank sends leaked-question alerts [5][2]. Any static item pool used at volume degrades on a schedule you do not control. TestGorilla separately lists VPNs, notes outside the webcam view, and deepfaked video or voice among the ways its own controls get bypassed [1].","introHtml":"Candidates use four broad methods, and only some of them leave any trace inside the assessment software.
\nA phone or second laptop beside the test window is the hardest method to detect, because nothing on that device touches the assessment session. TestGorilla, describing the limits of its own detection, notes that candidates can read from secondary devices or monitors the software does not track [1]. Webcam proctoring sees a face and a rough gaze direction, not what sits two feet to the left.
\nA second class of tool renders answers in a layer that never appears in a shared screen or recorded video feed. Cluely launched in 2025 around a hidden in-browser window the company called undetectable, prompting two startups to ship counter-detection products within weeks, and its chief executive answered by floating smart glasses or a recording necklace as ways past detection [11].
\nProxy test-taking and identity fraud are two of the four violation categories CodeSignal says its systems target [5]. The variant that defeats a webcam is a coached candidate, where the right person sits on camera while a helper watches a shared screen and feeds answers by phone.
\nCodeSignal runs a service called LeakSweep that monitors sites including Stack Overflow and LeetCode for leaked questions, and HackerRank sends leaked-question alerts [5][2]. Any static item pool used at volume degrades on a schedule you do not control. TestGorilla separately lists VPNs, notes outside the webcam view, and deepfaked video or voice among the ways its own controls get bypassed [1].
\n","outroMarkdown":null,"outroHtml":null,"contentJson":{},"configJson":{},"noteMarkdown":null,"noteHtml":null,"sortOrder":20},{"id":"d684ecd6-4f7a-4b4d-90ab-328c2937b083","sectionKey":"countermeasure_table","sectionType":"table_section","heading":"Which anti-cheat countermeasures catch what?","introMarkdown":"Every control below was checked against vendor documentation live in July 2026. Any detection rate here is a vendor claim, because no independent audit has been published.","introHtml":"Every control below was checked against vendor documentation live in July 2026. Any detection rate here is a vendor claim, because no independent audit has been published.
\n","outroMarkdown":"These fall into two groups: controls that watch the session, and controls that work on content and identity.","outroHtml":"These fall into two groups: controls that watch the session, and controls that work on content and identity.
\n","contentJson":{"rows":[{"cells":["Webcam and screen recording","Off-screen glances, another face in frame [1][5]","A phone outside the frame [1]","High","Video and biometric data [13][14]"]},{"cells":["Live human proctoring","Room setup and off-camera behavior [5]","Anything beyond camera view; costly at volume","Highest","Recording retention [13]"]},{"cells":["Lockdown browser","Other tabs and apps on the test machine [1]","Any second device; refused on personal hardware","Medium","Device access consent"]},{"cells":["Tab and focus tracking","Leaving the assessment window on that machine [1][2]","Phones; overlays that never take focus [11]","Low","Low"]},{"cells":["Keystroke and paste detection","Pasted blocks, irregular typing, code resets [1][3]","Answers retyped by hand from another screen [4]","Low","Low"]},{"cells":["Randomization and large item pools","Circulated answer keys and repeated items [1][5]","A model solving a fresh variant [7]","Low","Low"]},{"cells":["Code similarity and AI plagiarism models","Copied or structurally similar code, behavioral outliers [3][4]","Reprompted output; very short answers [3][4]","None","Flags are probabilistic"]},{"cells":["IP, device, and location checks","Two sessions from one address [1]","VPNs and residential proxies [1]","None","Location data handling"]},{"cells":["Biometric identity verification","Proxy test takers, multiple faces [2][5]","A helper off camera coaching the candidate","Medium","Illinois BIPA consent and damages [13]"]}],"columns":["Countermeasure","What it genuinely catches","What it misses","Candidate friction","Legal exposure"]},"configJson":{},"noteMarkdown":null,"noteHtml":null,"sortOrder":30},{"id":"3aa67567-a8fa-4f19-bfa5-e3d5ff204ab3","sectionKey":"session_controls","sectionType":"markdown_section","heading":"What do session monitoring controls actually catch?","introMarkdown":"Session monitoring watches the candidate and the test machine while the assessment runs, which makes it good at catching activity on that machine and blind to everything else.\n\n### Recorded and live proctoring\n\nRecorded proctoring covers the candidate's face, the room within the camera's field of view, and screen activity for the whole session [5]. Reviewing footage is real work, so set the review threshold before you turn it on, because a recording nobody watches is only a deterrent. A live proctor can ask a candidate to pan the room and stop a session in progress, which makes it the only control here that adapts to something unexpected [5], and the most expensive to run.\n\n### Lockdown browsers and tab tracking\n\nLocking the browser and logging tab or focus changes closes off the easiest route, which is opening a chatbot in the next tab [1][2]. Neither reaches a second device, and overlays that draw on top of the screen without taking window focus pass through both [11]. Candidates on personal machines also decline lockdown installs more often than teams expect.\n\n### Keystroke and paste detection\n\nPaste events, unusual typing rhythm, and large code resets separate original work from transcription, and HackerRank names code resetting and external copy-paste among the patterns its model watches [3]. A candidate who reads an answer off a phone and types it by hand produces a clean keystroke record, which is why HackerRank's own testing showed reworked model output passing similarity checks [4].","introHtml":"Session monitoring watches the candidate and the test machine while the assessment runs, which makes it good at catching activity on that machine and blind to everything else.
\nRecorded proctoring covers the candidate's face, the room within the camera's field of view, and screen activity for the whole session [5]. Reviewing footage is real work, so set the review threshold before you turn it on, because a recording nobody watches is only a deterrent. A live proctor can ask a candidate to pan the room and stop a session in progress, which makes it the only control here that adapts to something unexpected [5], and the most expensive to run.
\nLocking the browser and logging tab or focus changes closes off the easiest route, which is opening a chatbot in the next tab [1][2]. Neither reaches a second device, and overlays that draw on top of the screen without taking window focus pass through both [11]. Candidates on personal machines also decline lockdown installs more often than teams expect.
\nPaste events, unusual typing rhythm, and large code resets separate original work from transcription, and HackerRank names code resetting and external copy-paste among the patterns its model watches [3]. A candidate who reads an answer off a phone and types it by hand produces a clean keystroke record, which is why HackerRank's own testing showed reworked model output passing similarity checks [4].
\n","outroMarkdown":null,"outroHtml":null,"contentJson":{},"configJson":{},"noteMarkdown":null,"noteHtml":null,"sortOrder":40},{"id":"4cee25c0-b93a-4886-bc20-603c2e76bac4","sectionKey":"content_identity_controls","sectionType":"markdown_section","heading":"What do content and identity controls catch?","introMarkdown":"Content and identity controls work on the test material and on who is sitting for it, so they hold up better against a second device than session monitoring does.\n\n### Question randomization and large item pools\n\nRotating variants and retirement thresholds limit the damage when items leak, and CodeSignal rotates thousands of question variations so candidates do not receive identical assessments [5]. That solves the circulated-answer-key problem and does nothing about a model solving a fresh variant in under a minute [7]. Ask any vendor how many live items exist per skill and how often they retire.\n\n### Code similarity and AI plagiarism models\n\nSimilarity detection compares submissions against other candidates and known public solutions, and MOSS-style comparison survives cosmetic edits such as renamed variables. HackerRank's own 2023 testing showed that reprompting ChatGPT produced solutions that passed those checks cleanly, and the company put MOSS false positive rates as high as 70 percent [4]. Its replacement model combines code and behavioral signals at a documented 85 percent precision, with a caveat that detection is limited on very short answers [3].\n\n### Identity verification\n\nPhoto ID matching and multiple-face detection address the proxy test taker directly, which no other control here does [2][5]. It cannot see a helper sitting off camera, and it carries the most legal weight on the list, since face matching produces exactly the kind of data Illinois regulates [13].","introHtml":"Content and identity controls work on the test material and on who is sitting for it, so they hold up better against a second device than session monitoring does.
\nRotating variants and retirement thresholds limit the damage when items leak, and CodeSignal rotates thousands of question variations so candidates do not receive identical assessments [5]. That solves the circulated-answer-key problem and does nothing about a model solving a fresh variant in under a minute [7]. Ask any vendor how many live items exist per skill and how often they retire.
\nSimilarity detection compares submissions against other candidates and known public solutions, and MOSS-style comparison survives cosmetic edits such as renamed variables. HackerRank's own 2023 testing showed that reprompting ChatGPT produced solutions that passed those checks cleanly, and the company put MOSS false positive rates as high as 70 percent [4]. Its replacement model combines code and behavioral signals at a documented 85 percent precision, with a caveat that detection is limited on very short answers [3].
\nPhoto ID matching and multiple-face detection address the proxy test taker directly, which no other control here does [2][5]. It cannot see a helper sitting off camera, and it carries the most legal weight on the list, since face matching produces exactly the kind of data Illinois regulates [13].
\n","outroMarkdown":null,"outroHtml":null,"contentJson":{},"configJson":{},"noteMarkdown":null,"noteHtml":null,"sortOrder":50},{"id":"02308278-d63c-431c-804e-f40a77013ceb","sectionKey":"does_it_stop_ai","sectionType":"markdown_section","heading":"Does any of this stop ChatGPT?","introMarkdown":"No control on the market reliably stops a candidate consulting a chatbot on a device the assessment software cannot see. Detection raises the effort and the risk, because a candidate who pastes model output gets caught by paste logging and similarity checks while a candidate who retypes it usually does not [1][3][4].\n\nDetection of AI-assisted work rests on behavioral inference rather than on identifying machine-written text. HackerRank's model combines copy-paste activity, code similarity, time to solve, tab-switching, and iteration patterns, then returns a high or medium suspicion level at a stated 85 percent precision [3]. CodeSignal's Suspicion Score works the same way, turning solution similarity, telemetry, and copy-paste activity into a trust level per submission [5]. Both descriptions are the vendors' own, and neither has been audited publicly.\n\nThe setting that helps most removes the ambiguity, and HackerRank lets you declare and enforce whether candidates may use AI tools during an assessment [2]. Once AI use is either permitted and scored or banned and enforced, the assessment measures something you can name.\n\n### What anti-cheat technology is not\n\nIt is not proof. A flag is a probability, and at 85 percent precision roughly one flagged session in seven belongs to someone who did nothing wrong [3]. Rejecting on a flag alone, with no human review, converts a detection tool into a defensibility problem. It is also not a validity fix, since a test a model can solve measures the model.","introHtml":"No control on the market reliably stops a candidate consulting a chatbot on a device the assessment software cannot see. Detection raises the effort and the risk, because a candidate who pastes model output gets caught by paste logging and similarity checks while a candidate who retypes it usually does not [1][3][4].
\nDetection of AI-assisted work rests on behavioral inference rather than on identifying machine-written text. HackerRank's model combines copy-paste activity, code similarity, time to solve, tab-switching, and iteration patterns, then returns a high or medium suspicion level at a stated 85 percent precision [3]. CodeSignal's Suspicion Score works the same way, turning solution similarity, telemetry, and copy-paste activity into a trust level per submission [5]. Both descriptions are the vendors' own, and neither has been audited publicly.
\nThe setting that helps most removes the ambiguity, and HackerRank lets you declare and enforce whether candidates may use AI tools during an assessment [2]. Once AI use is either permitted and scored or banned and enforced, the assessment measures something you can name.
\nIt is not proof. A flag is a probability, and at 85 percent precision roughly one flagged session in seven belongs to someone who did nothing wrong [3]. Rejecting on a flag alone, with no human review, converts a detection tool into a defensibility problem. It is also not a validity fix, since a test a model can solve measures the model.
\n","outroMarkdown":null,"outroHtml":null,"contentJson":{},"configJson":{},"noteMarkdown":null,"noteHtml":null,"sortOrder":60},{"id":"2310928e-1b2e-452b-93f6-053157002ccf","sectionKey":"design_answer","sectionType":"markdown_section","heading":"Which assessment designs are hardest to outsource to a model?","introMarkdown":"Assessments built on your own material and followed by a live conversation are the hardest to outsource, and better evidence sits behind them than behind any monitoring stack. In the 2022 re-analysis of personnel selection validity, which corrected overcorrected range-restriction estimates and cut most validity coefficients by .10 to .20, structured interviews came out as the top-ranked selection procedure [10].\n\n### Work samples tied to something only you have\n\nA task drawn from your actual codebase or a scenario your team faced last quarter has no public solution for a model to retrieve and no answer key to circulate. HackerRank's multi-file project questions move in this direction, placing a candidate inside a realistic codebase instead of an isolated puzzle [2]. The cost is authoring effort, so most teams keep one or two per role family.\n\n### A live follow-up on submitted work\n\nA fifteen-minute conversation about a take-home separates the candidate who wrote it from the candidate who received it, with no surveillance at all. Ask why an approach was chosen and what would break at ten times the data volume. Someone who cannot narrate their own decisions has told you what you needed to know.\n\n### Tasks that require judgment about ambiguous context\n\nThe 2024 platform study shows where models weaken, since they were strong on archived, well-specified problems and much weaker on live contests [7]. Assessments built on incomplete requirements put the work back where a model gives least help.\n\n### Treat the assessment as a signal, not a verdict\n\nUsing a score to rank and route rather than to reject outright limits the damage a cheated result can do, and it softens the false positive problem, since a flag becomes a question for the next conversation rather than a silent rejection.","introHtml":"Assessments built on your own material and followed by a live conversation are the hardest to outsource, and better evidence sits behind them than behind any monitoring stack. In the 2022 re-analysis of personnel selection validity, which corrected overcorrected range-restriction estimates and cut most validity coefficients by .10 to .20, structured interviews came out as the top-ranked selection procedure [10].
\nA task drawn from your actual codebase or a scenario your team faced last quarter has no public solution for a model to retrieve and no answer key to circulate. HackerRank's multi-file project questions move in this direction, placing a candidate inside a realistic codebase instead of an isolated puzzle [2]. The cost is authoring effort, so most teams keep one or two per role family.
\nA fifteen-minute conversation about a take-home separates the candidate who wrote it from the candidate who received it, with no surveillance at all. Ask why an approach was chosen and what would break at ten times the data volume. Someone who cannot narrate their own decisions has told you what you needed to know.
\nThe 2024 platform study shows where models weaken, since they were strong on archived, well-specified problems and much weaker on live contests [7]. Assessments built on incomplete requirements put the work back where a model gives least help.
\nUsing a score to rank and route rather than to reject outright limits the damage a cheated result can do, and it softens the false positive problem, since a flag becomes a question for the next conversation rather than a silent rejection.
\n","outroMarkdown":null,"outroHtml":null,"contentJson":{},"configJson":{},"noteMarkdown":null,"noteHtml":null,"sortOrder":70},{"id":"0603939c-f46c-4892-880d-6a274f6a04a8","sectionKey":"tradeoffs_and_law","sectionType":"markdown_section","heading":"What do you give up by proctoring, and what is the legal exposure?","introMarkdown":"Proctoring costs completions and creates data you then have to protect and justify. Research on test takers found participants concerned about both the volume and the personal nature of information shared with proctoring companies [9]. A candidate with options who is asked to install monitoring software and record their room for a first-round screen has an easy alternative.\n\n### Biometric identity checks and Illinois BIPA\n\nFace matching produces a scan of face geometry, which the Illinois Biometric Information Privacy Act defines as a biometric identifier [13]. Before collecting one, a private entity must inform the subject in writing, state the purpose and length of storage, obtain a written release, and publish a retention schedule with destruction within three years of last interaction [13]. Section 20 gives any aggrieved person a private right of action with liquidated damages of $1,000 per negligent violation and $5,000 per intentional or reckless violation plus fees, which turns this into a litigation risk rather than paperwork [13].\n\n### Recorded video and the Illinois AI Video Interview Act\n\nAn employer that asks an applicant to record a video interview and uses AI to analyze it for an Illinois-based position must notify the applicant beforehand, explain how the AI works and what characteristics it evaluates, and obtain consent; on request the videos must be deleted within 30 days including backups [14]. An AI-scored recorded assessment with a video component sits close enough that counsel should decide, not the vendor.\n\n### Accommodation duties under the ADA\n\nAn employer must make reasonable accommodations, including in the administration of tests, to the known limitations of an otherwise qualified applicant with a disability, unless doing so is an undue hardship [12]. Proctoring rules collide with this regularly, since a screen reader triggers focus events and a movement disorder triggers gaze flags. Confirm before you buy that extended time and assistive technology run without generating integrity failures. This is general information about US federal and Illinois law rather than legal advice.","introHtml":"Proctoring costs completions and creates data you then have to protect and justify. Research on test takers found participants concerned about both the volume and the personal nature of information shared with proctoring companies [9]. A candidate with options who is asked to install monitoring software and record their room for a first-round screen has an easy alternative.
\nFace matching produces a scan of face geometry, which the Illinois Biometric Information Privacy Act defines as a biometric identifier [13]. Before collecting one, a private entity must inform the subject in writing, state the purpose and length of storage, obtain a written release, and publish a retention schedule with destruction within three years of last interaction [13]. Section 20 gives any aggrieved person a private right of action with liquidated damages of $1,000 per negligent violation and $5,000 per intentional or reckless violation plus fees, which turns this into a litigation risk rather than paperwork [13].
\nAn employer that asks an applicant to record a video interview and uses AI to analyze it for an Illinois-based position must notify the applicant beforehand, explain how the AI works and what characteristics it evaluates, and obtain consent; on request the videos must be deleted within 30 days including backups [14]. An AI-scored recorded assessment with a video component sits close enough that counsel should decide, not the vendor.
\nAn employer must make reasonable accommodations, including in the administration of tests, to the known limitations of an otherwise qualified applicant with a disability, unless doing so is an undue hardship [12]. Proctoring rules collide with this regularly, since a screen reader triggers focus events and a movement disorder triggers gaze flags. Confirm before you buy that extended time and assistive technology run without generating integrity failures. This is general information about US federal and Illinois law rather than legal advice.
\n","outroMarkdown":null,"outroHtml":null,"contentJson":{},"configJson":{},"noteMarkdown":null,"noteHtml":null,"sortOrder":80},{"id":"49106818-c662-4bb4-bfb6-fbba59f7ebb6","sectionKey":"contributor_perspective","sectionType":"markdown_section","heading":"How this answer was researched","introMarkdown":"This answer was assembled in July 2026 from vendor documentation opened live, published research on proctoring and model performance, and the primary statutory text on ilga.gov and eeoc.gov. No money changed hands with any assessment vendor. Two sourcing limits are worth stating plainly: every published detection accuracy figure in this field comes from the company selling the detection, and every prevalence figure counts flags rather than confirmed cheating, so both are labeled as vendor claims wherever they appear.\n\nIf you work in talent acquisition, assessment design, or industrial-organizational psychology, or you represent one of the platforms named here, corrections are welcome. An independent audit of detection accuracy, or a documented false positive rate, would make this record more accurate.","introHtml":"This answer was assembled in July 2026 from vendor documentation opened live, published research on proctoring and model performance, and the primary statutory text on ilga.gov and eeoc.gov. No money changed hands with any assessment vendor. Two sourcing limits are worth stating plainly: every published detection accuracy figure in this field comes from the company selling the detection, and every prevalence figure counts flags rather than confirmed cheating, so both are labeled as vendor claims wherever they appear.
\nIf you work in talent acquisition, assessment design, or industrial-organizational psychology, or you represent one of the platforms named here, corrections are welcome. An independent audit of detection accuracy, or a documented false positive rate, would make this record more accurate.
\n","outroMarkdown":null,"outroHtml":null,"contentJson":{},"configJson":{},"noteMarkdown":"This answer was written and reviewed by the AnswerStack Editorial Team, which has no commercial stake in the products, companies, or methods discussed. Every claim is cited inline and verified on the dates shown.","noteHtml":"This answer was written and reviewed by the AnswerStack Editorial Team, which has no commercial stake in the products, companies, or methods discussed. Every claim is cited inline and verified on the dates shown.
\n","sortOrder":90}],"citations":[{"title":"Cheating detection features in skills assessments","url":"https://www.testgorilla.com/blog/cheating-detection-skills-assessments/","excerpt":"Dual monitor setups make it harder to detect if candidates are viewing resources on a secondary screen.","quoteText":null,"sourceRole":"PRIMARY","verifiedAt":"2026-07-21T00:00:00","supportsText":"Named controls: IP address and location tracking, webcam and voice monitoring, full-screen monitoring, tab switch detection, time tracking, copy-paste detection, screenshot detection, developer tool detection, randomized questions, question exposure limits. Stated limits: detection can trigger false","domain":"testgorilla.com","publisherName":"TestGorilla"},{"title":"HackerRank Screen: Technical Assessments","url":"https://www.hackerrank.com/products/screen/","excerpt":"Clearly establish expectations for when candidates can and cannot use AI in your hiring process.","quoteText":null,"sourceRole":"PRIMARY","verifiedAt":"2026-07-21T00:00:00","supportsText":"Integrity stack: proprietary model for suspicious coding activity, AI proctor, identity verification with multiple-face detection, tab proctoring, copy and paste tracking, multiple-monitor detection, leaked question alerts, full session replay; configurable policy for whether candidates may use AI t","domain":"hackerrank.com","publisherName":"HackerRank"},{"title":"AI Plagiarism Detection","url":"https://support.hackerrank.com/articles/8000786908-ai-plagiarism-detection","excerpt":"The ML model currently has an overall precision of 85%, meaning flagged sessions are correct 85% of the time.","quoteText":null,"sourceRole":"PRIMARY","verifiedAt":"2026-07-21T00:00:00","supportsText":"Signals used by the ML model: external copy-pasting, structural code similarity and coding style, time taken to solve, tab-switching behavior, iteration counts, suspicious code resetting; results returned as High or Medium suspicion; stated overall precision of 85 percent meaning flagged sessions ar","domain":"support.hackerrank.com","publisherName":"HackerRank Knowledge Base"},{"title":"ChatGPT Easily Fools Traditional Plagiarism Detection","url":"https://www.hackerrank.com/blog/chatgpt-easily-fools-traditional-plagiarism-detection/","excerpt":"MOSS code similarity checks can be easily bypassed with ChatGPT.","quoteText":null,"sourceRole":"PRIMARY","verifiedAt":"2026-07-21T00:00:00","supportsText":"June 14, 2023 post demonstrating that reprompted ChatGPT solutions pass MOSS code similarity checks cleanly; vendor-stated figures that 25 percent of technical assessments show signs of plagiarism, that MOSS false positive rates run as high as 70 percent, and that HackerRank's AI-powered replacement","domain":"hackerrank.com","publisherName":"HackerRank"},{"title":"Cheating Detection and Fraud Prevention","url":"https://codesignal.com/cheating-and-fraud/","excerpt":"Rotate thousands of question variations so candidates don't receive identical assessments.","quoteText":null,"sourceRole":"PRIMARY","verifiedAt":"2026-07-21T00:00:00","supportsText":"Four violation categories: copy-paste plagiarism, proxy test-taking, unauthorized AI use, identity fraud. Named mechanisms: Suspicion Score combining solution similarity, telemetry and copy-paste activity; LeakSweep monitoring public sites including Stack Overflow and LeetCode for leaked questions; ","domain":"codesignal.com","publisherName":"CodeSignal"},{"title":"CodeSignal Detection Systems Identify and Stop Record-High Cheating Attempts as Assessment Fraud More Than Doubled in 2025","url":"https://www.prnewswire.com/news-releases/codesignal-detection-systems-identify-and-stop-record-high-cheating-attempts-as-assessment-fraud-more-than-doubled-in-2025-302696534.html","excerpt":"More than doubled, rising from 16 percent in 2024 to 35 percent.","quoteText":null,"sourceRole":"SUPPORTING","verifiedAt":"2026-07-21T00:00:00","supportsText":"Vendor-reported flagged fraud rate rising from 16 percent of assessments in 2024 to 35 percent in 2025; 40 percent on entry-level assessments; 48 percent in Asia Pacific and 27 percent in North America; flag breakdown of 35 percent off-screen referencing, 23 percent unusual typing with minimal pause","domain":"prnewswire.com","publisherName":"PR Newswire"},{"title":"Are Large Language Models a Threat to Programming Platforms? An Exploratory Study","url":"https://arxiv.org/html/2409.05824v1","excerpt":"Recruiters should supplement their evaluation methods with interviews, practical assessments, or real-world projects.","quoteText":null,"sourceRole":"INDEPENDENT","verifiedAt":"2026-07-21T00:00:00","supportsText":"ChatGPT-3.5, Gemini 1.0 Pro and Meta AI tested on 98 archived LeetCode problems, 126 archived Codeforces problems, 9 live contests and 2 HackerRank certification tests; LeetCode success of 71.43 percent for ChatGPT, 68.37 percent Gemini, 58.16 percent Meta AI; Codeforces 26.98 percent for ChatGPT; o","domain":"arxiv.org","publisherName":"arXiv"},{"title":"Effects of Proctoring on Online Intelligence Measurement: A Literature Overview and an Empirical Study","url":"https://pmc.ncbi.nlm.nih.gov/articles/PMC12470963/","excerpt":"Proctoring is crucial for easily cheatable tasks, such as memory tasks, but currently less critical for complex cognitive tasks.","quoteText":null,"sourceRole":"INDEPENDENT","verifiedAt":"2026-07-21T00:00:00","supportsText":"Cited meta-analysis of 49 studies and 100,434 participants reporting a standardized mean difference of 0.20 favoring unproctored scores; empirical study of 226 participants across proctored, unproctored random and unproctored chosen conditions finding no significant main effect of condition on score","domain":"pmc.ncbi.nlm.nih.gov","publisherName":"PubMed Central"},{"title":"Examining the Examiners: Students' Privacy and Security Perceptions of Online Proctoring Services","url":"https://arxiv.org/abs/2106.05917","excerpt":"Participants are concerned about both the amount and the personal nature of the information shared with the exam proctoring companies.","quoteText":null,"sourceRole":"INDEPENDENT","verifiedAt":"2026-07-21T00:00:00","supportsText":"Analysis of proctoring browser extension reviews plus an online survey of 102 test takers; participants concerned about both the amount and the personal nature of information shared with proctoring companies; institutional power dynamics and trust shaping willingness to accept monitoring","domain":"arxiv.org","publisherName":"arXiv"},{"title":"Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range","url":"https://pubmed.ncbi.nlm.nih.gov/34968080/","excerpt":"Most of the same selection procedures that ranked high in prior summaries remain high in rank, but with mean validity estimates reduced by .10-.20 points.","quoteText":null,"sourceRole":"INDEPENDENT","verifiedAt":"2026-07-21T00:00:00","supportsText":"Sackett, Zhang, Berry and Lievens, 2022; correction of systematic overcorrection for range restriction reduces mean validity estimates by .10 to .20; structured interviews emerge as the top-ranked selection procedure; predictor-criterion relationships considerably lower than previously thought","domain":"pubmed.ncbi.nlm.nih.gov","publisherName":"Journal of Applied Psychology (Sackett, Zhang, Berry & Lievens, 2022) via PubMed"},{"title":"Startups launch products to catch people using AI cheating app Cluely","url":"https://techcrunch.com/2025/04/29/startups-launch-products-to-catch-people-using-ai-cheating-app-cluely/","excerpt":"Whether it's smart glasses, a transparent glass screen overlay, a recording necklace, or even a brain chip, we're not sure.","quoteText":null,"sourceRole":"INDEPENDENT","verifiedAt":"2026-07-21T00:00:00","supportsText":"Cluely's hidden in-browser window marketed as undetectable; counter-detection products launched by Validia (Truely) and Proctaroo within weeks; Cluely chief executive naming smart glasses, a transparent glass screen overlay, a recording necklace or a brain chip as possible ways past detection","domain":"techcrunch.com","publisherName":"TechCrunch"},{"title":"Employment Tests and Selection Procedures","url":"https://www.eeoc.gov/laws/guidance/employment-tests-and-selection-procedures","excerpt":"Fail to make reasonable accommodations, including in the administration of tests, to the known physical or mental limitations of an otherwise qualified individual with a disability.","quoteText":null,"sourceRole":"PRIMARY","verifiedAt":"2026-07-21T00:00:00","supportsText":"ADA obligation to make reasonable accommodations, including in the administration of tests, to the known physical or mental limitations of an otherwise qualified individual with a disability unless it imposes undue hardship; Title VII requirement that a selection procedure with disparate impact be j","domain":"eeoc.gov","publisherName":"U.S. Equal Employment Opportunity Commission"},{"title":"740 ILCS 14, Biometric Information Privacy Act (full text)","url":"https://www.ilga.gov/Legislation/ILCS/Articles?ActID=3004&ChapterID=57&Print=True","excerpt":"Against a private entity that negligently violates a provision of this Act, liquidated damages of $1,000 or actual damages, whichever is greater.","quoteText":null,"sourceRole":"PRIMARY","verifiedAt":"2026-07-21T00:00:00","supportsText":"Section 10 definition of biometric identifier including a scan of hand or face geometry; Section 15(a) written retention policy and destruction when the purpose is satisfied or within 3 years of last interaction; Section 15(b) written notice of collection, written notice of specific purpose and leng","domain":"ilga.gov","publisherName":"Illinois General Assembly"},{"title":"Artificial Intelligence Video Interview Act, 820 ILCS 42","url":"https://www.ilga.gov/Legislation/ILCS/Articles?ActID=4015&ChapterID=68&Print=True","excerpt":"Obtain, before the interview, consent from the applicant to be evaluated by the artificial intelligence program.","quoteText":null,"sourceRole":"PRIMARY","verifiedAt":"2026-07-21T00:00:00","supportsText":"Employers using AI to analyze applicant video interviews for Illinois-based positions must notify the applicant before the interview, explain how the AI works and what general types of characteristics it uses, and obtain consent before the interview; sharing limited to persons whose expertise or tec","domain":"ilga.gov","publisherName":"Illinois General Assembly"}],"revisions":[],"relatedAnswers":[{"id":"72492b35-0af1-4a64-9156-e4f70d3a5618","slug":"ai-resume-screening-software-actually-work","question":"Does AI resume screening software actually work, and is it worth it (bias, accuracy)?","publishedAt":"2026-07-21T21:20:44.609","confidenceScore":82,"confidenceLabel":"High","industry":{"id":"e1a1da3c-52a8-4e99-88ad-d08ef17d7c3f","slug":"hr-technology","label":"HR Technology","description":"ATS, HRIS, payroll, and workforce management"},"topic":{"slug":"candidate-screening-assessment","label":"Candidate Screening & Assessment","description":"Background Checks, Pre-Employment Testing, Reference Checks, Skills Assessment","schemaKind":null},"contributor":{"id":"ec39deab-44fe-48d8-9029-fefe993ab85a","slug":"answer-stack","displayName":"AnswerStack","websiteUrl":null},"snippet":"Independent audits find language-model resume rankers favor some demographic groups on otherwise identical resumes. The deeper issue is what these tools are trained to predict, usually a past hiring decision rather than job performance, so a high accuracy score can mean only that the model agrees with last year's recruiters. AI screening holds up on high-volume roles with concrete requirements, used to rank rather than reject, with human review and monitored adverse impact.","url":"/q/ai-resume-screening-software-actually-work"},{"id":"7c1df232-4afc-4f90-acc6-373bb41cf94d","slug":"screening-background-check-software-integrate-with-my-ats","question":"Does screening / background check software integrate with my ATS or HRIS?","publishedAt":"2026-07-21T21:20:40.075","confidenceScore":88,"confidenceLabel":"High","industry":{"id":"e1a1da3c-52a8-4e99-88ad-d08ef17d7c3f","slug":"hr-technology","label":"HR Technology","description":"ATS, HRIS, payroll, and workforce management"},"topic":{"slug":"candidate-screening-assessment","label":"Candidate Screening & Assessment","description":"Background Checks, Pre-Employment Testing, Reference Checks, Skills Assessment","schemaKind":null},"contributor":{"id":"ec39deab-44fe-48d8-9029-fefe993ab85a","slug":"answer-stack","displayName":"AnswerStack","websiteUrl":null},"snippet":"Screening and assessment vendors ship certified ATS and HRIS connectors, so ordering, status, and results move without rekeying. The mechanics that decide whether the integration helps are narrower: where the FCRA disclosure and authorization are hosted, what writes back, who can read results, and whether adverse action fires from the screening platform or your ATS.","url":"/q/screening-background-check-software-integrate-with-my-ats"},{"id":"ef2314e9-45b8-4867-adcb-bb8bb029b6e7","slug":"pre-employment-assessments-and-background-checks-legal-eeoc-ada-fcra","question":"Are pre-employment assessments and background checks legal (EEOC, ADA, FCRA, disparate impact)?","publishedAt":"2026-07-21T21:20:37.515","confidenceScore":86,"confidenceLabel":"High","industry":{"id":"e1a1da3c-52a8-4e99-88ad-d08ef17d7c3f","slug":"hr-technology","label":"HR Technology","description":"ATS, HRIS, payroll, and workforce management"},"topic":{"slug":"candidate-screening-assessment","label":"Candidate Screening & Assessment","description":"Background Checks, Pre-Employment Testing, Reference Checks, Skills Assessment","schemaKind":null},"contributor":{"id":"ec39deab-44fe-48d8-9029-fefe993ab85a","slug":"answer-stack","displayName":"AnswerStack","websiteUrl":null},"snippet":"Screening is lawful in principle and regulated in the details. Third-party background reports trigger FCRA disclosure, authorization, and a two-step adverse action sequence. Tests and criminal-record screens face Title VII adverse impact analysis under the Uniform Guidelines. The ADA restricts what you may ask or examine before a conditional offer. Fair-chance, salary history, credit, and automated hiring tool rules vary by state and city. A June 2026 federal opinion narrowed enforcement posture on disparate impact without repealing the statute.","url":"/q/pre-employment-assessments-and-background-checks-legal-eeoc-ada-fcra"},{"id":"c0112a94-3c83-4520-9777-fff00c999af6","slug":"pre-employment-tests-skills-assessments-actually-worth-it-do-they","question":"Are pre-employment tests and skills assessments actually worth it, and do they predict job performance?","publishedAt":"2026-07-21T21:20:35.197","confidenceScore":88,"confidenceLabel":"High","industry":{"id":"e1a1da3c-52a8-4e99-88ad-d08ef17d7c3f","slug":"hr-technology","label":"HR Technology","description":"ATS, HRIS, payroll, and workforce management"},"topic":{"slug":"candidate-screening-assessment","label":"Candidate Screening & Assessment","description":"Background Checks, Pre-Employment Testing, Reference Checks, Skills Assessment","schemaKind":null},"contributor":{"id":"ec39deab-44fe-48d8-9029-fefe993ab85a","slug":"answer-stack","displayName":"AnswerStack","websiteUrl":null},"snippet":"Selection tests do predict performance, but the evidence base was revised downward in 2022 after researchers showed that decades of range restriction corrections had inflated published validity. Structured interviews now top the ranking at .42, while general cognitive ability sits at .31 and carries the largest Black-White subgroup difference of any common method. Value depends on volume, measurable job requirements, what the assessment replaces, and whether you can monitor selection rates by group.","url":"/q/pre-employment-tests-skills-assessments-actually-worth-it-do-they"}],"contributorStats":{"verifiedAnswers":224,"openDisputes":0},"schemaJson":{"@context":"https://schema.org","@type":"Question","name":"Do candidates cheat on online skills assessments, and does proctoring / anti-cheat actually stop it (including AI/ChatGPT)?","text":"Do candidates cheat on online skills assessments, and does proctoring / anti-cheat actually stop it (including AI/ChatGPT)?","url":"https://www.answerstack.io/q/candidates-cheat-on-online-skills-assessments-and-does-proctoring","answerCount":1,"datePublished":"2026-07-21T21:20:42.255","author":{"@type":"Person","name":"AnswerStack Editorial Team","worksFor":{"@type":"Organization","name":"AnswerStack"},"url":"https://www.answerstack.io/contributors/answer-stack"},"about":[{"@type":"Thing","name":"Candidate Screening & Assessment"},{"@type":"Thing","name":"HR Technology"}],"acceptedAnswer":{"@type":"Answer","text":"Candidates do cheat on online skills assessments, and generative AI made the most common formats easier to outsource, particularly coding challenges and knowledge tests. A 2024 study found ChatGPT-3.5 solved 71.43 percent of 98 archived LeetCode problems, and all three models tested earned HackerRank problem-solving certificates at basic and intermediate levels [7]. Proctoring narrows the gap without closing it, since the meta-analysis most cited in this area, covering 49 studies and 100,434 test takers, put unproctored scores ahead by a standardized mean difference of only 0.20 [8], and vendors say plainly that a candidate reading from a second monitor or phone is the case their software cannot see [1]. Detection accuracy is vendor-published, including HackerRank's stated 85 percent precision on its AI plagiarism model, which means roughly one flagged session in seven is wrong [3]. Assessment design carries better evidence than surveillance, since structured interviews ranked as the top selection procedure in the 2022 re-analysis of personnel selection validity [10].","url":"https://www.answerstack.io/q/candidates-cheat-on-online-skills-assessments-and-does-proctoring","upvoteCount":0,"datePublished":"2026-07-21T21:20:42.255","dateModified":"2026-07-21T00:00:00","author":{"@type":"Person","name":"AnswerStack Editorial Team","worksFor":{"@type":"Organization","name":"AnswerStack"},"url":"https://www.answerstack.io/contributors/answer-stack"},"citation":[{"@type":"CreativeWork","name":"Cheating detection features in skills assessments","url":"https://www.testgorilla.com/blog/cheating-detection-skills-assessments/"},{"@type":"CreativeWork","name":"HackerRank Screen: Technical Assessments","url":"https://www.hackerrank.com/products/screen/"},{"@type":"CreativeWork","name":"AI Plagiarism Detection","url":"https://support.hackerrank.com/articles/8000786908-ai-plagiarism-detection"},{"@type":"CreativeWork","name":"ChatGPT Easily Fools Traditional Plagiarism Detection","url":"https://www.hackerrank.com/blog/chatgpt-easily-fools-traditional-plagiarism-detection/"},{"@type":"CreativeWork","name":"Cheating Detection and Fraud Prevention","url":"https://codesignal.com/cheating-and-fraud/"},{"@type":"CreativeWork","name":"CodeSignal Detection Systems Identify and Stop Record-High Cheating Attempts as Assessment Fraud More Than Doubled in 2025","url":"https://www.prnewswire.com/news-releases/codesignal-detection-systems-identify-and-stop-record-high-cheating-attempts-as-assessment-fraud-more-than-doubled-in-2025-302696534.html"},{"@type":"CreativeWork","name":"Are Large Language Models a Threat to Programming Platforms? An Exploratory Study","url":"https://arxiv.org/html/2409.05824v1"},{"@type":"CreativeWork","name":"Effects of Proctoring on Online Intelligence Measurement: A Literature Overview and an Empirical Study","url":"https://pmc.ncbi.nlm.nih.gov/articles/PMC12470963/"},{"@type":"CreativeWork","name":"Examining the Examiners: Students' Privacy and Security Perceptions of Online Proctoring Services","url":"https://arxiv.org/abs/2106.05917"},{"@type":"CreativeWork","name":"Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range","url":"https://pubmed.ncbi.nlm.nih.gov/34968080/"},{"@type":"CreativeWork","name":"Startups launch products to catch people using AI cheating app Cluely","url":"https://techcrunch.com/2025/04/29/startups-launch-products-to-catch-people-using-ai-cheating-app-cluely/"},{"@type":"CreativeWork","name":"Employment Tests and Selection Procedures","url":"https://www.eeoc.gov/laws/guidance/employment-tests-and-selection-procedures"},{"@type":"CreativeWork","name":"740 ILCS 14, Biometric Information Privacy Act (full text)","url":"https://www.ilga.gov/Legislation/ILCS/Articles?ActID=3004&ChapterID=57&Print=True"},{"@type":"CreativeWork","name":"Artificial Intelligence Video Interview Act, 820 ILCS 42","url":"https://www.ilga.gov/Legislation/ILCS/Articles?ActID=4015&ChapterID=68&Print=True"}]}}}