Clearview AI
| Company type | Private |
|---|---|
| Industry | Facial recognition technology |
| Founded | 2017 |
| Founder | Hoan Ton-That, Charles C. Johnson, Richard Schwartz |
| Headquarters | Manhattan , United States |
| Products | Facial recognition software |
Clearview AI, Inc. is a U.S.-based company specializing in facial recognition technology, primarily serving law enforcement and governmental bodies. Its system compares facial images against a database containing more than 20 billion pictures collected from online sources, including social media platforms. Established in 2017 by Hoan Ton-That, Charles C. Johnson, and Richard Schwartz, the firm stayed largely under the radar until late 2019 when its adoption by police agencies became publicly known.
The technology developed by Clearview AI has ignited significant debate, with privacy concerns voiced by U.S. senators and legal action initiated by the American Civil Liberties Union (ACLU) over alleged privacy infringements. Nevertheless, various U.S. law enforcement agencies have utilized the software to identify and detain suspects. The company's operations have led to penalties and prohibitions within European Union countries for violating privacy regulations, alongside ongoing investigations both domestically and internationally. In 2022, Clearview reached a settlement with the ACLU, agreeing to restrict the sale of its facial recognition products within the United States exclusively to government organizations.
A data breach in 2020 exposed that approximately 2,200 entities across 27 nations had accounts granting access to Clearview’s facial recognition tools.
History
Clearview AI was launched in 2017 by Hoan Ton-That and Richard Schwartz after transferring assets from their prior venture SmartCheckr, which they co-founded with Charles C. Johnson. The company was incorporated in Manhattan following their encounter at the Manhattan Institute. Initial capital raised amounted to $8.4 million from investors such as Kirenaga Partners and Peter Thiel. In 2020, Clearview secured an additional $8.625 million during a subsequent funding round with undisclosed backers, followed by a $30 million investment round in 2021. Early versions of Clearview’s application were made available to prospective Series A investors, including billionaire John Catsimatidis, who tested it in his New York grocery stores as a tool for preventing theft.
By October 2020, Clearview AI's valuation exceeded $100 million. In May 2021, the company expanded its executive team by appointing its first chief strategy officer, chief revenue officer, and chief marketing officer: Devesh Ashra (formerly deputy assistant secretary at the U.S. Treasury), Chris Metaxas (previously with LexisNexis Risk Solutions), and Susan Crandall (ex-marketing executive at LexisNexis and Motorola Solutions). Both Ashra and Metaxas left the firm later that same year.
In August 2021, Clearview established an advisory board including prominent individuals such as Raymond Kelly, Richard A. Clarke, and Floyd Abrams. By October 2021, the company reported having collected over 10 billion images through web scraping techniques. In May 2022, Clearview broadened its market reach by offering its facial recognition software to educational institutions and international lending organizations.
To protect its business model amid various legal challenges, Clearview assembled a distinguished legal team featuring attorneys like Tor Ekeland, Lee S. Wolosky, Paul Clement, Floyd Abrams, and Jack Mulcaire. Abrams emphasized that the conflict between privacy rights and free speech could potentially be adjudicated by the Supreme Court.
On February 19, 2025, following Hoan Ton-That’s resignation in December 2024, Clearview appointed early backers Hal Lambert and Richard Swartz as co-chief executives. Lambert had previously overseen fundraising efforts for Donald Trump’s presidential campaigns. Ton-That continues to serve on the company’s board of directors.
Usage
Clearview AI provides facial recognition technology that enables users to upload a photograph of a face and compare it against a vast database, offering links to online appearances of the matched individual. The company maintained a low profile until a 2020 investigative article by The New York Times exposed its operations, revealing tactics such as falsifying corporate information and deleting founders’ social media profiles. After this exposure, more than 40 technology and civil rights organizations called on the Privacy and Civil Liberties Oversight Board (PCLOB) and lawmakers to halt the use of facial recognition due to concerns over bias and privacy violations.
The emergence of Clearview ignited an international discussion on the governance of facial recognition tools in policing. Law enforcement officials commended its enhanced capacity to identify suspects from multiple viewpoints compared to older systems. Conversely, leading platforms including Twitter, Google, YouTube, Facebook, and Venmo issued cease-and-desist orders demanding Clearview cease scraping images from their sites, citing breaches of their policies. Founder Hoan Ton-That defended these actions under First Amendment protections, stating that the company had collected over 50 billion images.
In New Zealand, police initiated a trial of Clearview’s system following an approach by Marko Jukic in early 2020, suggesting it might have assisted in identifying the Christchurch mosque shooter. However, this trial attracted criticism for lacking authorization from senior police officials or the Privacy Commissioner. Similarly, during the May 2020 protests in Miami sparked by George Floyd’s death, Clearview was employed to identify a protest participant; local media coverage and company statements emphasized that its use was for investigative purposes rather than mass surveillance.
In late 2020, the ACLU of Washington urged the Seattle Police Department to halt use of Clearview after discovering officers accessed the system without official permission. The city auditor noted that while existing regulations addressed sanctioned surveillance activities, they did not cover unauthorized individual usage.
Following the January 6 Capitol attack, law enforcement in Oxford, Alabama utilized Clearview to analyze images released by the FBI of suspects involved, forwarding pertinent leads to federal agencies. By March 2022, Ukraine’s Ministry of Defence adopted Clearview technology to identify Russian soldiers, counter misinformation, and document casualties using more than two billion images sourced from Russian social media platforms. By April 2022, Ukrainian authorities had conducted over 5,000 searches with Clearview’s assistance, which was provided free of charge.
Defense attorneys have also employed Clearview’s capabilities; notably in Florida, it aided in locating a witness that contributed to dismissal of vehicular homicide charges. Use by U.S. law enforcement has surged rapidly—over one million searches were logged in 2022, with numbers doubling in 2023.
Marketing efforts and pushback
Clearview AI encouraged adoption by offering free trials directly to individual law enforcement officers instead of entire agencies. The company leveraged strong ties with the Republican Party to engage police departments. Initial onboarding emails urged users to perform up to 100 searches. In 2020, Clearview priced its facial recognition software at roughly one-tenth the cost of competitors.
Clearview asserted that its technology played a role in apprehending a terrorist via the New York Police Department tip line and claimed involvement in solving other cases in New York as well as 40 cold cases by submitting identifications through tip lines. Although NYPD officially denied institutional partnerships, reports indicated that thirty officers possessed Clearview accounts in 2020. Subsequent documents revealed collaboration between Clearview and NYPD leadership along with a vendor contract—raising concerns about unregulated officer access.
The New Jersey Attorney General issued a cease-and-desist order after Clearview used his likeness in promotional materials. Despite confirming use of the software in identifying a child predator case, New Jersey banned police use of Clearview altogether. The marketing video was swiftly removed following objections.
In March 2020, amid the COVID-19 pandemic, Clearview proposed using its technology for contact tracing efforts aimed at tracking virus exposure across states. This initiative drew criticism from U.S. senators and privacy advocates who warned against exploiting the health crisis as a pretext for intrusive surveillance deployment. Tests demonstrated that Clearview could recognize individuals even when faces were partially obscured by masks.
A data breach later that year revealed Clearview’s clientele extended beyond law enforcement to include private security firms and casinos. Plans for international expansion included sales targeting authoritarian regimes.
Senator Edward J. Markey voiced concerns about Clearview’s potential threat to personal anonymity and requested comprehensive business disclosures regarding privacy safeguards, bias mitigation, and security practices. After Clearview declined these requests, Markey sent additional letters expressing worry about sales to authoritarian governments and potential violations of COPPA. He criticized attempts to justify expanded surveillance during the pandemic.
Senator Ron Wyden also raised privacy issues and reported multiple canceled meetings with Clearview’s founder.
Despite ongoing controversies, Time recognized Clearview AI among its list of the 100 most influential companies in 2021.
Accuracy
In October 2021, Clearview presented its facial recognition algorithm for evaluation in one of the two accuracy assessments conducted by the National Institute of Standards and Technology (NIST). Among 300 algorithms tested on identifying two images of the same individual, Clearview ranked within the top 10. Nevertheless, it did not participate in NIST’s test designed to match an unknown face against a database containing 10 billion images, a scenario more representative of its practical application. This marked the initial third-party examination of the software.
During 2020, Clearview reported internal accuracy figures ranging from 98.6% to 100%, though these results were derived from tests lacking diverse or representative population samples.
In 2021, Clearview revealed plans to develop "deblur" and "mask removal" technologies aimed at improving the clarity of obscured or blurred facial images through machine learning techniques. The company acknowledged that these enhancements might increase error rates and stated they would be utilized exclusively to generate investigative leads rather than definitive identifications.
Miami Assistant Chief of Police Armando Aguilar noted in 2023 that Clearview’s AI had assisted in resolving multiple homicide investigations, with approximately 450 uses per year. He stressed that matches produced by the system function only as investigative leads, with arrests depending on traditional police work.
Although documented instances of misidentification exist, limited transparency regarding law enforcement usage suggests the actual frequency could be higher. Founder Hoan Ton-That asserts near-perfect accuracy when using mugshots but admits that the quality of input images significantly influences outcomes.
Data breaches
In February 2020, Clearview AI experienced a data breach that exposed its list of clients. The company’s legal representative, Tor Ekeland, confirmed that the vulnerability was promptly addressed. Following this incident, the United States House Committee on Science, Space, and Technology sought detailed information about Clearview’s biometric data handling and security protocols.
Despite Clearview’s app being intended for private customer use only, its Android and iOS versions were discovered stored in unsecured Amazon S3 buckets. Documentation uncovered detailed steps for installing the app via an enterprise certificate, circumventing the App Store. As a result, Apple revoked Clearview’s developer privileges due to terms violations, disabling the application. Analysis of the Android app’s source code revealed integration with Google Play Services, requests for precise geolocation data, voice search functionality, augmented reality features using Vuzix hardware, and barcode scanning capabilities for driver’s licenses.
In April 2020, cybersecurity firm SpiderSilk’s Mossab Hussein identified exposed Clearview source code repositories caused by misconfigured user permissions. These repositories contained sensitive credentials including secret keys for cloud storage and Slack, as well as compiled applications and pre-release builds. Although Hussein reported these findings to Clearview, he declined to sign their bug bounty non-disclosure agreement. Founder Ton-That accused Hussein of attempting extortion. Furthermore, Hussein discovered a bucket containing 70,000 videos from a Rudin Management residential property.
Insight Camera
Clearview runs an additional venture called Insight Camera that provides AI-powered security camera systems aimed at sectors such as retail, banking, and residential properties. Prominent clients include the United Federation of Teachers and Rudin Management. Following BuzzFeed’s investigation linking Insight Camera to Clearview AI, its website was taken offline.
Customer base
A data breach disclosed that Clearview maintained approximately 2,200 active clients spanning 27 countries. Among these, at least 600 are law enforcement agencies, predominantly located in the United States and Canada, with ongoing international growth. Although the company's primary focus is on serving police departments, it also holds contracts with corporations such as Bank of America, Kohl's, Macy's, along with several academic institutions.
In September 2025, Clearview secured a federal contract valued at $10 million—the largest U.S. federal agreement to date—surpassing its previous $2.3 million contract awarded in 2021 by U.S. Immigration and Customs Enforcement (ICE).
- U.S. law enforcement and government
- Global law enforcement
Litigation
Clearview AI has encountered numerous legal disputes across different regions due to its operational model. The company has both defended itself in court, reached settlements in some instances, and withdrawn from certain markets.
Clearview’s assertion of a First Amendment protection over publicly available information has been challenged by privacy authorities such as Scott Skinner-Thompson and Margot Kaminski, who emphasize risks related to continuous surveillance and loss of anonymity. Former New York City Police Commissioner Bill Bratton recognized privacy concerns but supported rigorous law enforcement standards.
United States
After a New York Times exposé in January 2020, states including Illinois, California, Virginia, and New York initiated lawsuits alleging infringements on privacy rights. Most cases were consolidated within New York’s Southern District Court. Additional litigation arose in Vermont and Illinois; the latter invoked the Biometric Information Privacy Act (BIPA), which prohibits companies from using residents’ biometric data without permission. Clearview contended that Illinois statutes did not apply due to its New York-based operations.
In May 2020, to address BIPA-related issues, Clearview stopped servicing non-government clients and blocked photo access from Illinois locations. It also terminated accounts belonging to private customers and organizations situated in Illinois. Following this, the ACLU filed a class action lawsuit under BIPA provisions.
By June 2024, Clearview settled by transferring 23% equity—valued at $52 million—instead of a cash settlement that could have jeopardized the company’s financial stability.
In May 2022, Clearview resolved the ACLU lawsuit by agreeing to limit access to its facial recognition database exclusively to law enforcement agencies.
In Vermont, Clearview’s defense invoking Section 230 immunity was denied because the case centered on how it collected photographs rather than content created by third parties.
Canada
Clearview AI withdrew from the Canadian market in July 2020 amid probes into its police usage. Daniel Therrien, Canada’s Privacy Commissioner, condemned the firm’s practice of collecting biometric data without consent as unlawful mass surveillance. In June 2021, Therrien determined that the Royal Canadian Mounted Police had breached privacy laws through hundreds of unauthorized searches using Clearview AI.
European Union and United Kingdom
In January 2021, Hamburg’s Data Protection Authority (DPA) ruled Clearview AI’s biometric database unlawful under GDPR regulations and mandated deletion of biometric data for an individual despite Clearview lacking EU offices. The European Center for Digital Rights (NOYB) criticized this limited scope of enforcement. Earlier in March 2020, Hamburg DPA requested Clearview’s client list due to data protection compliance requirements.
By May 2021, legal complaints were lodged against Clearview in Austria, France, Greece, Italy, and the UK for breaches of European privacy statutes concerning online data collection. In November 2021, the UK’s Information Commissioner’s Office (ICO) issued a preliminary order demanding cessation of processing data belonging to UK residents over multiple violations. At that time, Clearview denied operating or having customers within the UK. In May 2023, BBC reported that Clearview was fined over £7.5 million by the UK privacy regulator and instructed to erase facial recognition data of UK citizens—marking its fourth significant penalty after Australia, France, and Italy. However, in October 2023 this fine was overturned following an appeal challenging ICO’s authority over actions conducted by foreign governments.
In September 2024, the Dutch Data Protection Authority imposed a €30.5 million fine on Clearview for unlawfully gathering facial images without consent—constituting GDPR violations due to the intrusive nature of biometric processing and insufficient transparency regarding use of such data.
On October 7, 2025, London's Upper Tribunal affirmed that GDPR applies to Clearview AI and upheld enforcement actions along with monetary penalties issued by the ICO in May 2022.