Skip to content
FaceSearch Wiki
Try it

Timeline

Face recognition and face search, year by year

Face recognition and face search from the first experiments to today, event by event.

Events per decade
  • 1960s 1
  • 1990s 2
  • 2000s 3
  • 2010s 10
  • 2020s 27

Source: This timeline

1960s

  1. January 1, 1964 · Technology

    Woodrow Bledsoe's team builds the first semi-automatic face matching system

    At Panoramic Research in Palo Alto, Woodrow Bledsoe, Helen Chan Wolf and Charles Bisson built a system in which an operator marked facial landmarks on photographs with a tablet and a computer compared the measured distances against a database. Funded by an unnamed intelligence agency, the work was little published at the time. The exact start is not documented to the month; 1964 to 1965 is the period usually cited, and the first of January is used here as a placeholder. It established the pattern every later system followed: reduce a face to measurements, then compare numbers.

    How a face search worksFace recognition Source: Wikipedia, Woody Bledsoe

1990s

  1. January 1, 1991 · Research

    Turk and Pentland publish Eigenfaces for Recognition

    Matthew Turk and Alex Pentland at the MIT Media Lab showed that faces could be represented as combinations of a small set of 'eigenfaces' derived by principal component analysis, and recognised by projecting a new image into that space. Published in the Journal of Cognitive Neuroscience in 1991 (presented at CVPR the same year), the method ran in near real time on 1991 hardware and became the baseline against which face recognition was measured for a decade. The month is not fixed to a single date here. It introduced the idea of a compact numeric face representation that modern embeddings descend from.

    Face embeddingEmbeddings and similarity Source: Turk and Pentland, Eigenfaces for Recognition (PDF)

  2. September 1, 1993 · Research

    The FERET programme starts systematic face recognition testing

    The Face Recognition Technology (FERET) programme, sponsored by the US Department of Defense's Counterdrug Technology Development Program and run with NIST, began collecting a standard image database and running comparable tests across research systems from 1993. It produced the first large benchmark (over 14,000 images) and the evaluation habits that NIST's later Face Recognition Vendor Tests inherited. The programme start is dated to September 1993 in NIST's account. FERET made it possible to say which algorithm was better on the same data, a question that previously had no answer.

    National Institute of Standards and Technology (NIST)Accuracy, bias and testing Source: NIST, Face Recognition Technology (FERET)

2000s

  1. December 1, 2001 · Research

    Viola and Jones publish the real-time face detector

    Paul Viola and Michael Jones presented a face detection method at CVPR 2001 that combined simple rectangular features, an integral image for fast computation, AdaBoost feature selection and a cascade of classifiers. It detected faces in video at frame rate on 2001 hardware, which no previous method had done. Face detection, finding where a face is before anything else can happen, became cheap enough to ship in cameras and later phones. The December date reflects the conference proceedings; the paper was presented earlier in the year.

    Face detectionFace detection and alignment Source: IEEE, Rapid object detection using a boosted cascade of simple features

  2. October 1, 2007 · Research

    Labeled Faces in the Wild is released

    Researchers at the University of Massachusetts Amherst released Labeled Faces in the Wild (LFW), more than 13,000 face images of about 5,700 people collected from news photographs on the web, with pairs labelled as same or different person. It became the standard benchmark for face verification in unconstrained conditions and the dataset on which the deep-learning results of 2014 and 2015 were reported. The release is dated to October 2007 by the technical report. Later work noted that LFW's demographics skew towards white male public figures, a limitation that shaped the bias debate a decade on.

    Accuracy, bias and testingFalse positive Source: UMass Amherst, Labeled Faces in the Wild

  3. January 1, 2008 · Technology

    Social networks make billions of face photos public

    By 2008 Facebook was receiving tens of millions of photo uploads a day and Flickr had passed two billion images, most with faces, many public by default. The flood of labelled, public face photos is the raw material that made both the 2014 deep-learning breakthroughs and the later web-scale face indexes possible. No single event marks the change, so the first of January 2008 is used as a placeholder for a shift that took place across the decade's second half. The datasets entry covers how research collections were later built from these sources.

    Building a web-scale face indexPublic web Source: Facebook newsroom

2010s

  1. December 15, 2010 · Product

    Facebook introduces face recognition tag suggestions

    Facebook announced tag suggestions, which used face recognition to propose the names of friends in newly uploaded photos, on 15 December 2010 for US users and expanded it in 2011. It was the first face recognition feature used daily by hundreds of millions of people and, because it was on by default, the first to draw regulator attention at scale: Ireland's Data Protection Commissioner audit and complaints in Germany followed in 2011 and 2012, and the feature was switched off in the EU in 2012.

    General Data Protection Regulation (GDPR) and biometric data/timeline/ Source: Facebook newsroom, Making Photo Tagging Easier

  2. June 24, 2014 · Research

    Facebook's DeepFace reaches near-human accuracy on LFW

    At CVPR 2014, Facebook researchers presented DeepFace, a nine-layer deep network trained on four million labelled faces that reached 97.35 percent on the LFW verification benchmark, within a fraction of a point of the human figure reported for the same test. It marked the switch from hand-designed features to learned representations and showed what web-scale labelled photo collections could do. The conference ran 24 to 27 June 2014. From this point, accuracy gains came from data and compute rather than from new feature design.

    Face embeddingEmbeddings and similarity Source: Meta Research, DeepFace: Closing the Gap to Human-Level Performance in Face Verification

  3. April 1, 2015 · Case

    Illinois users sue Facebook under BIPA over tag suggestions

    In April 2015, Illinois residents filed class actions alleging that Facebook's tag suggestions created face templates without the written consent that the state's Biometric Information Privacy Act requires. Consolidated in federal court in California as In re Facebook Biometric Information Privacy Litigation, the case survived dismissal, was certified as a class in 2018 and ended in a settlement of 650 million dollars approved in 2021. It made BIPA the most consequential biometric statute in the United States. The first filing is dated to April 2015; the exact day is not fixed here.

    Illinois Biometric Information Privacy Act (BIPA)Facebook BIPA settlement (USD 650 million) Source: Wikipedia, Biometric Information Privacy Act

  4. June 7, 2015 · Research

    Google's FaceNet introduces the embedding approach used by modern face search

    FaceNet, presented at CVPR 2015 (7 to 12 June), trained a network to map each face directly to a 128-dimensional vector such that distances between vectors correspond to face similarity, using a triplet loss. It reported 99.63 percent on LFW. The method is the template for how face search engines work today: compute an embedding for every face found on the web, store the vectors, and answer a query by nearest-neighbour search. The paper was posted to arXiv in March 2015.

    Face embeddingSimilarity scoreBuilding a web-scale face index Source: arXiv, FaceNet: A Unified Embedding for Face Recognition and Clustering

  5. May 17, 2016 · Product

    FindFace lets anyone identify strangers on VK

    FindFace, built on NTechLab's algorithm, launched in early 2016 and let users photograph a stranger and find their VKontakte profile with about 70 percent accuracy, according to the company. Within months it had been used to identify and harass women who appeared in adult films, a case widely reported in May 2016. The public app was withdrawn in 2018 as the company moved to enterprise and government contracts. It was the first mass demonstration of consumer face search against a social network and a preview of the debate that followed PimEyes and Clearview.

    Search4FacesOne-to-many matching (1:N) Source: The Guardian, Face recognition app taking Russia by storm may bring end to public anonymity

  6. January 1, 2017 · Product

    PimEyes launches as a public face search engine

    PimEyes was started in Poland in 2017 by two developers as a face search engine over publicly accessible web pages: upload a face, get the pages where a similar face appears. It later moved its registered operations to the Seychelles and changed ownership. Through 2020 to 2022 it became the engine most cited in press and regulator discussion of consumer face search. The launch year is documented; the month is not, so the first of January is used as a placeholder.

    PimEyesReverse face search Source: netzpolitik.org, PimEyes investigation

  7. February 1, 2018 · Research

    Gender Shades documents accuracy gaps by skin type and gender

    Joy Buolamwini and Timnit Gebru's Gender Shades study, presented at the Conference on Fairness, Accountability and Transparency in February 2018, tested three commercial gender classification systems on a dataset balanced by skin type and gender. Error rates for darker-skinned women reached up to 34.7 percent, against under 1 percent for lighter-skinned men. The study is about gender classification, not identification, but it set the terms of the bias debate and led vendors to re-evaluate training data. It is the reference point for later demographic testing by NIST.

    Accuracy, bias and testingNational Institute of Standards and Technology (NIST) Source: Gender Shades project site

  8. May 25, 2018 · Law

    The GDPR starts to apply across the EU, classing biometric data as special category

    Regulation (EU) 2016/679 became applicable on 25 May 2018. Article 9 lists biometric data processed to uniquely identify a person among the special categories whose processing is prohibited unless a specific exception applies, and Article 3 extends the rules to companies outside the EU that monitor people in it. Every later European action against face search engines, from the Clearview fines to the PimEyes proceedings, rests on these two articles. The regulation also gave individuals the erasure right that the opt-out routes of several engines try to accommodate.

    General Data Protection Regulation (GDPR) and biometric dataBiometric data Source: EUR-Lex, Regulation (EU) 2016/679

  9. January 25, 2019 · Case

    Illinois Supreme Court: no actual harm needed to sue under BIPA

    In Rosenbach v. Six Flags Entertainment Corp., decided 25 January 2019, the Illinois Supreme Court held that a person whose biometric data is collected without the notice and consent BIPA requires is 'aggrieved' and may sue without showing further injury. The case concerned a teenager's fingerprint taken for a season pass. The ruling removed the main defence companies had relied on and opened the way for the volume of BIPA litigation that followed, including against face recognition features.

    Illinois Biometric Information Privacy Act (BIPA)Rosenbach v. Six Flags Entertainment Corp. Source: Illinois Courts, Rosenbach v. Six Flags opinion

  10. December 19, 2019 · Research

    NIST publishes its demographic effects study of face recognition algorithms

    NIST Interagency Report 8280, published 19 December 2019, evaluated 189 algorithms from 99 developers and found that false positive rates varied by demographic group by factors of 10 to 100 for many algorithms, with higher rates for West and East African and East Asian faces in one-to-one matching, and elevated false positives for Black women in some one-to-many searches. It also found that the most accurate algorithms showed much smaller differences. The report is the largest independent test of demographic effects and is cited by regulators and courts on both sides of the accuracy debate.

    Accuracy, bias and testingNational Institute of Standards and Technology (NIST)False positive Source: NIST, NISTIR 8280 Face Recognition Vendor Test Part 3: Demographic Effects

2020s

  1. January 18, 2020 · Incident

    The New York Times reveals Clearview AI's three-billion-image face database

    Kashmir Hill's report on 18 January 2020 disclosed that Clearview AI had scraped more than three billion photos from social media and other sites to build a face search tool sold to law enforcement, and that hundreds of agencies were already using it. The story triggered cease-and-desist letters from Twitter, Google and Facebook, investigations by regulators in Canada, Australia, the UK, France, Italy, Greece and the Netherlands, and the ACLU's lawsuit under BIPA. Nearly every case entry on this wiki concerning Clearview begins here.

    Clearview AIACLU v Clearview AI (Illinois BIPA settlement)Information Commissioner v Clearview AI (United Kingdom) Source: The New York Times, The Secretive Company That Might End Privacy as We Know It

  2. May 28, 2020 · Case

    The ACLU sues Clearview AI under Illinois BIPA

    On 28 May 2020 the American Civil Liberties Union and partner organisations filed suit in Illinois state court alleging that Clearview AI had captured and stored face templates of Illinois residents without consent, in violation of BIPA. The case settled in 2022 with Clearview agreeing to a permanent nationwide ban on selling its database to most private companies and a five-year ban on access by Illinois government bodies. It remains the most significant private enforcement against a face search company in the United States.

    ACLU v Clearview AI (Illinois BIPA settlement)Illinois Biometric Information Privacy Act (BIPA)Clearview AI Source: ACLU, ACLU Sues Clearview AI

  3. July 10, 2020 · Incident

    netzpolitik.org investigates PimEyes and finds private individuals in its results

    On 10 July 2020 the German outlet netzpolitik.org published an investigation showing that PimEyes returned matches for ordinary people, not only public figures, and that its index included photos from blogs, forums and small sites across Europe. The report quoted the company's claim to index hundreds of millions of faces and prompted the Baden-Wurttemberg data protection authority to examine the service. It is the earliest detailed public account of a consumer face search engine's reach in Europe.

    PimEyesBaden-Wurttemberg data protection authority v PimEyes (Germany) Source: netzpolitik.org, Gesichter-Suchmaschine PimEyes schafft Anonymitat ab

  4. January 1, 2021 · Case

    Baden-Wurttemberg's data protection authority opens proceedings against PimEyes

    Following the 2020 investigations, the Landesbeauftragte fur den Datenschutz und die Informationsfreiheit Baden-Wurttemberg, the German state authority responsible because of PimEyes' then presence, opened administrative proceedings examining whether the engine's processing of biometric data of people in Germany complied with the GDPR. The proceedings continued after the company relocated to the Seychelles and raised the jurisdiction questions that later arose with Clearview. The opening date is not published to the day; 2021 is used with the first of January as a placeholder.

    Baden-Wurttemberg data protection authority v PimEyes (Germany)PimEyesGeneral Data Protection Regulation (GDPR) and biometric data Source: LfDI Baden-Wurttemberg

  5. February 3, 2021 · Case

    Canadian privacy commissioners find Clearview AI's scraping unlawful

    On 3 February 2021 the Office of the Privacy Commissioner of Canada and the commissioners of Quebec, British Columbia and Alberta published joint findings that Clearview AI's collection of images of Canadians without consent was mass surveillance and a violation of federal and provincial privacy law. Clearview had stopped offering its service in Canada in July 2020 but disputed the findings. The provincial authorities issued binding orders in December 2021 requiring deletion of Canadians' data. The findings are the first full regulator determination against a face search database.

    Privacy Commissioners v Clearview AI (Canada)Canada: PIPEDA, the federal private-sector privacy lawOffice of the Privacy Commissioner of Canada (OPC) Source: Office of the Privacy Commissioner of Canada, PIPEDA Findings 2021-001

  6. October 14, 2021 · Case

    Australia's privacy regulator finds Clearview AI breached the Privacy Act

    On 14 October 2021 the Office of the Australian Information Commissioner determined that Clearview AI had breached the Privacy Act 1988 by collecting Australians' sensitive biometric information without consent, by unfair means, and without taking reasonable steps to notify them or ensure accuracy. The Commissioner ordered Clearview to stop collecting and to destroy images of people in Australia. Clearview sought review; the OAIC announced in August 2024 that it would not pursue further proceedings while the orders stood. The determination is the Australian anchor for every later face search question.

    OAIC v Clearview AI (Australia)Australia: Privacy Act 1988Office of the Australian Information Commissioner (OAIC) Source: OAIC, Clearview AI breached Australians' privacy

  7. November 2, 2021 · Product

    Meta shuts down Facebook's face recognition system and deletes a billion templates

    On 2 November 2021 Meta announced that Facebook would stop using face recognition for tag suggestions and automatic alt text and would delete the face templates of more than a billion people. The company cited growing societal concern and the absence of clear rules. The decision came after the 650 million dollar BIPA settlement and during regulatory scrutiny in several countries. It removed the largest consumer face recognition deployment and showed that a platform could reverse the 2010 default.

    Facebook BIPA settlement (USD 650 million)Illinois Biometric Information Privacy Act (BIPA) Source: Meta newsroom, An Update On Our Use of Face Recognition

  8. March 9, 2022 · Case

    Italy's Garante fines Clearview AI EUR 20 million and bans further processing

    On 9 March 2022 the Italian data protection authority, the Garante, fined Clearview AI 20 million euros for processing biometric and geolocation data of people in Italy without a legal basis, ordered the deletion of the data and banned further collection. The decision followed complaints by privacy groups and the authority's own investigation. It was the first of three EUR 20 million fines issued by EU authorities in 2022; France and Greece followed within months.

    Garante v Clearview AI (Italy)General Data Protection Regulation (GDPR) and biometric dataClearview AI Source: Garante per la protezione dei dati personali, decision of 10 February 2022 (docweb 9751362)

  9. May 9, 2022 · Case

    Clearview AI settles the ACLU BIPA case with a ban on most private sales

    On 9 May 2022 Clearview AI and the ACLU announced a settlement of the 2020 Illinois lawsuit. Clearview agreed to a permanent nationwide ban on selling or giving access to its face database to most private businesses and individuals, a five-year ban on access by Illinois state and local government bodies, and an opt-out form for Illinois residents. The settlement did not include an admission of liability. It fixed the business model of the company at law-enforcement and government customers in the United States.

    ACLU v Clearview AI (Illinois BIPA settlement)Clearview AIIllinois Biometric Information Privacy Act (BIPA) Source: ACLU, In Big Win, Settlement Ensures Clearview AI Complies With Groundbreaking Illinois Biometric Privacy Law

  10. May 23, 2022 · Case

    The UK ICO fines Clearview AI GBP 7.5 million and orders deletion of UK residents' data

    On 23 May 2022 the Information Commissioner's Office issued a monetary penalty of 7,552,800 pounds and an enforcement notice requiring Clearview AI to stop obtaining and using the personal data of UK residents and to delete it. The ICO found that Clearview had no lawful basis, had not met the standards for special category data, and had not told people their images were collected. Clearview appealed on jurisdiction grounds, which led to the First-tier Tribunal decision of 2023 and the Upper Tribunal judgment of 2025.

    Information Commissioner v Clearview AI (United Kingdom)Information Commissioner's Office (ICO)UK GDPR and the Data Protection Act 2018 Source: ICO, ICO fines facial recognition database company Clearview AI Inc

  11. May 26, 2022 · Incident

    The New York Times tests PimEyes and calls it alarmingly accurate

    Kashmir Hill's article of 26 May 2022 described testing PimEyes with photographs of consenting journalists and finding decades-old images, conference photos and pictures the subjects did not know existed. It documented the pricing, the opt-out that requires an ID, and the company's position that the service is for searching one's own face with no mechanism to enforce it. The piece brought PimEyes to a mainstream audience and is the most cited account of what a consumer face search engine finds.

    PimEyesOpt-out Source: The New York Times, A Face Search Engine Anyone Can Use Is Alarmingly Accurate

  12. October 20, 2022 · Case

    France's CNIL fines Clearview AI EUR 20 million

    On 20 October 2022 the CNIL imposed the maximum fine of 20 million euros on Clearview AI for unlawful processing of biometric data, failure to honour access and erasure requests, and failure to cooperate with the authority, after a formal notice in 2021 went unanswered. The CNIL ordered Clearview to stop collecting data on people in France and to delete what it held, with a penalty of 100,000 euros per day of delay after two months. In 2023 the CNIL added a 5.2 million euro penalty for non-compliance.

    CNIL v Clearview AI (France)CNIL (Commission nationale de l'informatique et des libertés)General Data Protection Regulation (GDPR) and biometric data Source: CNIL, Facial recognition: 20 million euros penalty against Clearview AI

  13. February 17, 2023 · Case

    Illinois Supreme Court: each scan is a separate BIPA violation

    In Cothron v. White Castle System, decided 17 February 2023, the Illinois Supreme Court held that a BIPA claim accrues each time biometric data is collected or disclosed without consent, not only the first time. The case concerned fingerprint scans used for employee computer access over years. The decision raised potential damages to figures the court itself called potentially crippling and prompted the 2024 amendment limiting recovery to one violation per person per method of collection.

    Cothron v. White Castle System, Inc.Illinois Biometric Information Privacy Act (BIPA) Source: Illinois Courts, Cothron v. White Castle System opinion

  14. October 17, 2023 · Case

    UK First-tier Tribunal allows Clearview's appeal on jurisdiction

    On 17 October 2023 the First-tier Tribunal held that the ICO lacked jurisdiction to fine Clearview AI because the company's processing was carried out on behalf of foreign law enforcement clients and fell outside the UK GDPR's material scope. The ICO announced in November 2023 that it would seek permission to appeal. The decision was reversed in part by the Upper Tribunal in October 2025, which found the processing was within scope; Clearview was granted permission to take the point to the Court of Appeal.

    Information Commissioner v Clearview AI (United Kingdom)Information Commissioner's Office (ICO) Source: ICO, Information Commissioner seeks permission to appeal Clearview AI Inc ruling

  15. December 19, 2023 · Case

    The FTC bans Rite Aid from using face recognition for five years

    On 19 December 2023 the Federal Trade Commission announced a proposed order barring Rite Aid from using facial recognition technology for surveillance for five years, after finding the pharmacy chain had deployed it in hundreds of stores without reasonable safeguards, generating false matches that led to shoppers, disproportionately women and people of colour, being wrongly accused of shoplifting. The order is the FTC's first substantial action on face recognition use by a retailer and sets out the accuracy testing and notice the agency expects.

    FTC v Rite Aid (facial recognition surveillance order)Accuracy, bias and testing Source: FTC, Rite Aid Banned from Using AI Facial Recognition After FTC Says Retailer Deployed Technology without Reasonable Safeguards

  16. January 1, 2024 · Case

    Belgium's Data Protection Authority orders measures against PimEyes

    During 2024 the Belgian Data Protection Authority (Gegevensbeschermingsautoriteit) issued an order against PimEyes concerning the processing of biometric data of people in Belgium without a legal basis, following a complaint. The decision is one of two European regulator actions against the engine documented on this wiki, alongside the Baden-Wurttemberg proceedings. The month of the decision is not fixed in the sources used here, so the year is given with the first of January as a placeholder; the case entry carries the detail.

    PimEyes: Belgian and Hamburg regulators and the noyb lawsuitPimEyes Source: Belgian Data Protection Authority

  17. August 1, 2024 · Law

    The EU AI Act enters into force, with a ban on untargeted scraping of facial images

    Regulation (EU) 2024/1689, the Artificial Intelligence Act, entered into force on 1 August 2024. Article 5 prohibits placing on the market or using AI systems that create or expand facial recognition databases through untargeted scraping of facial images from the internet or CCTV footage, a provision written with Clearview-style databases in mind. The prohibitions became applicable on 2 February 2025. The Act also classes remote biometric identification systems as high risk and restricts real-time use in public spaces by law enforcement.

    EU Artificial Intelligence Act and the ban on scraping facesClearview AICrawler Source: EUR-Lex, Regulation (EU) 2024/1689

  18. September 3, 2024 · Case

    The Dutch Data Protection Authority fines Clearview AI EUR 30.5 million

    On 3 September 2024 the Autoriteit Persoonsgegevens announced a fine of 30.5 million euros against Clearview AI for building an illegal database of over 30 billion images, including Dutch residents, and for failing to honour access and erasure rights, with further penalty payments of up to 5.1 million euros for non-compliance. The authority also said it was examining whether directors could be held personally liable. It is the largest fine against a face search company to date.

    Autoriteit Persoonsgegevens v Clearview AI (Netherlands)General Data Protection Regulation (GDPR) and biometric dataClearview AI Source: JURIST, Netherlands data protection authority fines US AI company 30.5M euros over facial recognition database

  19. February 2, 2025 · Law

    The AI Act's prohibitions, including the facial scraping ban, become applicable

    From 2 February 2025 the practices listed in Article 5 of the AI Act are prohibited across the EU, including the untargeted scraping of facial images to build or expand face recognition databases. Penalties for prohibited practices run up to 35 million euros or 7 percent of global turnover. The provision applies to any provider or deployer whose system is placed on the EU market or whose output is used in the EU, which brings foreign face search engines within reach if they serve EU users.

    EU Artificial Intelligence Act and the ban on scraping facesBuilding a web-scale face index Source: EUR-Lex, Regulation (EU) 2024/1689, Article 113 (application dates)

  20. May 19, 2025 · Law

    The TAKE IT DOWN Act is signed into US federal law

    On 19 May 2025 the President signed S.146, the TAKE IT DOWN Act, which makes it a federal crime to knowingly publish non-consensual intimate images, including realistic AI-generated ones, and requires covered platforms to establish a notice-and-removal process and remove reported images within 48 hours. Platforms had one year to comply, with enforcement by the Federal Trade Commission from May 2026. It is the first federal statute in the United States addressing synthetic intimate imagery.

    TAKE IT DOWN Act (United States)NCII (non-consensual intimate images)Removing intimate images shared without consent, by country Source: Congress.gov, S.146 TAKE IT DOWN Act

  21. October 8, 2025 · Case

    The UK Upper Tribunal rules that the ICO had jurisdiction over Clearview AI

    On 8 October 2025 the Upper Tribunal (Administrative Appeals Chamber) upheld three of the Information Commissioner's four grounds of appeal, finding that Clearview AI's processing related to monitoring the behaviour of UK residents and was within the scope of the UK GDPR, and sent the case back to the First-tier Tribunal to decide the substantive appeal against the 2022 fine. In December 2025 the Upper Tribunal granted Clearview permission to appeal to the Court of Appeal. The judgment is the current statement of how far the UK regime reaches foreign face search operators.

    Information Commissioner v Clearview AI (United Kingdom)Information Commissioner's Office (ICO)UK GDPR and the Data Protection Act 2018 Source: ICO, UK Upper Tribunal hands down judgment on Clearview AI Inc

  22. November 14, 2025 · Law

    India notifies the Digital Personal Data Protection Rules, 2025

    On 14 November 2025 the Ministry of Electronics and Information Technology notified the rules that give operational effect to the Digital Personal Data Protection Act, 2023, including the consent notice requirements, the Data Protection Board's procedures, breach notification and the category of Significant Data Fiduciaries for large-scale or sensitive processing. The rules phase in over 18 months. For face search, the Act's consent-based model and the Board's complaint route are the first statutory tools Indian residents have against the processing of their facial data by private services.

    India's Digital Personal Data Protection Act, 2023Indian Cyber Crime Coordination Centre (I4C) Source: Ministry of Electronics and Information Technology, Government of India

  23. January 1, 2026 · Incident

    Indian police bust matrimonial fraud rackets using AI-generated 'brides'

    In 2026 police in Durg, Chhattisgarh, arrested the operator of a fake matrimonial call centre that used AI-generated photographs of women who did not exist, and Delhi Police dismantled a pan-India racket run from Agra that used fake profiles and staged calls with women posing as prospective brides, with victims estimated in the thousands. The cases illustrate a shift from stolen photos, which a face search can trace, to synthetic faces, which have no source to find. The month is not fixed here; the year is documented.

    CatfishingDeepfakeDeepfakes and synthetic faces Source: ThePrint, Fake brides, staged calls: Delhi Police busts pan-India matrimonial fraud racket

  24. May 1, 2026 · Product

    YouTube opens likeness detection to all adult creators

    In May 2026 YouTube extended its likeness detection tool, first offered to a pilot group of public figures and creators, to every creator over 18. Creators enrol with a selfie-style scan, after which YouTube flags uploaded videos that appear to use their face, including AI-generated ones, and offers a removal request. It is the first face-matching tool offered by a major platform for the benefit of the person depicted rather than for tagging or advertising. The month is documented; the exact rollout day varied by region, so the first of May is used.

    DeepfakeDeepfakes and synthetic faces Source: YouTube Official Blog

  25. May 19, 2026 · Law

    The FTC begins enforcing the TAKE IT DOWN Act's 48-hour removal duty

    On 19 May 2026, one year after signature, the Federal Trade Commission announced that it had begun enforcing the platform provisions of the TAKE IT DOWN Act and launched TakeItDown.ftc.gov for victims whose removal requests were ignored. Covered platforms must remove reported non-consensual intimate images, and known identical copies, within 48 hours. The same month the FTC sent warning letters to twelve companies offering 'nudify' tools that lacked a removal process. Civil penalties run up to 53,088 dollars per violation.

    TAKE IT DOWN Act (United States)Removing intimate images shared without consent, by countryDeepfake Source: FTC, FTC Begins Enforcing the TAKE IT DOWN Act

  26. November 1, 2026 · Law

    The EDPB warns on smart glasses paired with face recognition

    In November 2026 the European Data Protection Board adopted a preliminary statement on the data protection risks of smart glasses, warning that the devices collect personal data in a near-invisible way and that pairing them with facial or voice recognition in public spaces is high risk under the GDPR. The statement followed consumer launches of camera glasses and reports of hobbyist projects linking them to face search engines. The day of adoption is not fixed in the sources used here; the first of November is used as a placeholder.

    European Data Protection Board (EDPB)General Data Protection Regulation (GDPR) and biometric dataOne-to-many matching (1:N) Source: European Data Protection Board

  27. November 4, 2026 · Law

    UK minister promises legislation on police use of live facial recognition

    On 4 November 2026 the UK Crime and Policing Minister announced that the government would legislate for police use of facial recognition in public places, with a legal framework, oversight and proportionality safeguards, after years of deployment under common-law powers and the 2020 Bridges v South Wales Police ruling that found the existing framework deficient. The announcement concerns law enforcement rather than consumer face search, but the same debate about biometric data in public spaces shapes both.

    UK GDPR and the Data Protection Act 2018Information Commissioner's Office (ICO) Source: Biometric Update, UK unveils plan for police facial recognition guardrails and oversight