PhotoDNA

From HandWiki

PhotoDNA is perceptual hashing algorithm for image-identification and content filtering technology[1] widely used by online service providers.[2][3] Although the design of the algorithm has never been publicly disclosed by its authors, a leaked software library that implements it became accessible through a repository on GitHub in 2021.[4]

History

PhotoDNA was developed by Microsoft Research and Hany Farid, professor at Dartmouth College, beginning in 2009. From a database of known images and video files, it creates unique hashes to represent each image, which can then be used to identify other instances of those images.[5]

The hashing method initially relied on converting images into a black-and-white format, dividing them into squares, and quantifying the shading of the squares,[6] did not employ facial recognition technology, nor could it identify a person or object in the image. The method sought to be resistant to alterations in the image, including resizing and minor color alterations.[5] Since 2015,[7] similar methods are used for individual video frames in video files.[8]

Microsoft donated the PhotoDNA technology to Project VIC, managed and supported by the International Centre for Missing & Exploited Children (ICMEC) and used as part of digital forensics operations[9][10] by storing "fingerprints" that can be used to uniquely identify an individual photo.[10][11] The database includes hashes for millions of items.[12]

In December 2014, Microsoft made PhotoDNA available to qualified organizations in a software as a service model for free through the Azure Marketplace.[13]

In the 2010s and 2020s, PhotoDNA was put forward in connection with policy proposals relating to content moderation and internet censorship,[14] including US Senate hearings (2019 on "digital responsibility",[2] 2022 on the EARN IT Act[15]) and various proposals by the European Commission dubbed "upload filters" by civil society[16][17] such as so-called voluntary codes (in 2016[18] on hate speech[19] after 2015 events, 2018[20] and 2022[21] on disinformation), copyright legislation (chiefly the 2019 copyright directive debated between 2014[22] and 2021[23]), terrorism-related regulations (TERREG)[24] and internet wiretapping regulations (2021 "chat control").[25]

In 2016, Hany Farid proposed to extend usage of the technology to terrorism-related content.[26] In December 2016, Facebook, Twitter, Google and Microsoft announced plans to use PhotoDNA to remove extremist content such as terrorist recruitment videos or violent terrorist imagery.[27] In 2018 Facebook stated that PhotoDNA was used to automatically remove al-Qaeda videos.[14]

By 2019, big tech companies including Microsoft, Facebook and Google publicly announced that since 2017 they were running the GIFCT as a shared database of content to be automatically censored.[2] As of 2021, Apple was thought to be using NeuralHash for similar purposes.[28]

In 2022, The New York Times covered the story of two dads whose Google accounts were closed after photos they took of their child for medical purposes were automatically uploaded to Google's servers.[29] The article compares PhotoDNA, which requires a database of known hashes, with Google's AI-based technology, which can recognize previously unseen exploitative images. [30][31]

Usage

As a hashing algorithm, PhotoDNA is now used for multiple types of image-based online abuse including CSAM detection and detection of non-consensual intimate imagery.[32] Microsoft originally used PhotoDNA on its own services including Bing and OneDrive.[33] As of 2022, PhotoDNA was widely used by online service providers for their content moderation efforts[11][34][35] including Google's Gmail, Twitter,[36] Facebook,[37] Adobe Systems,[38] Reddit,[39] and Discord.[40]

The UK Internet Watch Foundation, which has been compiling a reference database of PhotoDNA signatures, reportedly had over 300,000 hashes of known child sexual exploitation materials.[41] The National Center for Missing & Exploited Children (NCMEC) maintains hash databases using the PhotoDNA format as well as others such as PDQ, MD5, and SHA-1.[42][43][44]

PhotoDNA is widely used to identify known illegal content, disable accounts, and report individuals to investigative organizations.[8]

Access

Access to PhotoDNA is limited due to the sensitive nature of the technology. Online platforms and law enforcement agencies can apply for access through Microsoft, or through organizations allowed to sublicense the technology. The Technology Coalition provides access for members and non-member online platforms, and Kindred Tech offers PhotoDNA for law enforcement agencies.[45][46]

See also

References

  1. ↑ Douze, Matthijs; Tolias, Giorgos; Pizzi, Ed; Papakipos, Zoë; Chanussot, Lowik; Radenovic, Filip; Jenicek, Tomas; Maximov, Maxim; Leal-Taixé, Laura; Elezi, Ismail; Chum, Ondřej; Ferrer, Cristian Canton (2022-02-21). "The 2021 Image Similarity Dataset and Challenge". arXiv:2106.09672 [cs.CV]. Image fingerprints, such as PhotoDNA from Microsoft, are used throughout the industry to identify images that depict child exploitation and abuse
  2. ↑ 2.0 2.1 2.2 "The Rise of Content Cartels". 2020-02-11. http://knightcolumbia.org/content/the-rise-of-content-cartels. 
  3. ↑ Hill, Kashmir (2022-08-21). "A Dad Took Photos of His Naked Toddler for the Doctor. Google Flagged Him as a Criminal.". The New York Times. ISSN 0362-4331. https://www.nytimes.com/2022/08/21/technology/google-surveillance-toddler-photo.html. 
  4. ↑ Deryck, Maxime; Leblanc-Albarel, Diane; Preneel, Bart (2026), White-Box Attacks on PhotoDNA Perceptual Hash Function, 2026/486, https://eprint.iacr.org/2026/486, retrieved 2026-03-15 
  5. ↑ 5.0 5.1 "New Technology Fights Child Porn by Tracking Its "PhotoDNA"". Microsoft Corporation. December 15, 2009. https://news.microsoft.com/2009/12/15/new-technology-fights-child-porn-by-tracking-its-photodna/#sm.0001mpmupctevct7pjn11vtwrw6xj. 
  6. ↑ "Photo DNA: Step by step". Microsoft. http://www.microsoft.com/global/en-us/news/publishingimages/ImageGallery/Images/Infographics/PhotoDNA/flowchart_photodna_Web.jpg. 
  7. ↑ "How PhotoDNA for Video is being used to fight online child exploitation". September 12, 2018. https://news.microsoft.com/on-the-issues/2018/09/12/how-photodna-for-video-is-being-used-to-fight-online-child-exploitation/. 
  8. ↑ 8.0 8.1 "How PhotoDNA for Video is being used to fight online child exploitation". news.microsoft.com. September 12, 2018. https://news.microsoft.com/on-the-issues/2018/09/12/how-photodna-for-video-is-being-used-to-fight-online-child-exploitation/. 
  9. ↑ Jackson, William (August 27, 2014). "Improved image analysis tools speed exploited children cases". GCN. http://gcn.com/articles/2014/08/27/image-analysis-exploited-children.aspx. 
  10. ↑ 10.0 10.1 Clark, Liat (April 30, 2014). "Child abuse-tracking tech donated to the world". Wired UK. https://www.wired.co.uk/news/archive/2014-04/30/video-fingerprints-child-abuse. 
  11. ↑ 11.0 11.1 "Microsoft's response to the consultation on the European Commission Communication on the Rights of the Child (2011–2014)". http://ec.europa.eu/justice/news/consulting_public/0009/contributions/registered_organisations/099_microsoft.pdf. , European Commission
  12. ↑ Ward, Mark (March 23, 2014). "Cloud-based archive tool to help catch child abusers". BBC News. https://www.bbc.com/news/technology-26612059. 
  13. ↑ "PhotoDNA Cloud Service". Microsoft Corporation. http://www.microsoft.com/en-us/photodna. 
  14. ↑ 14.0 14.1 Richard Allan (2018-06-18). "Hearing at 11:14". http://www.europarl.europa.eu/website/webstreaming.html?event=20180619-0900-COMMITTEE-IMCO.  in "The EU's horizontal regulatory framework for illegal content removal in the DSM". http://www.europarl.europa.eu/committees/en/events-hearings.html?id=20180613CHE04321. 
  15. ↑ Thu; Szoka, Feb 10th 2022 03:30pm-Berin; Cohn, Ari (2022-02-10). "The Top Ten Mistakes Senators Made During Today's EARN IT Markup". https://www.techdirt.com/2022/02/10/top-ten-mistakes-senators-made-during-todays-earn-it-markup/. 
  16. ↑ Schmon, Christoph (2021-06-03). "The EU Commission's Refusal to Let Go of Filters". https://www.eff.org/deeplinks/2021/06/eu-commissions-guidance-article-17-did-not-let-go-filters. 
  17. ↑ "Upload filters: a danger to free internet content?". March 28, 2019. https://www.ionos.com/digitalguide/websites/digital-law/upload-filters/. 
  18. ↑ "Fighting illegal online hate speech: first assessment of the new code of conduct". 2016-12-06. https://ec.europa.eu/newsroom/just/items/50840. 
  19. ↑ "The EU Code of conduct on countering illegal hate speech online | European Commission". Ec.europa.eu. https://ec.europa.eu/info/policies/justice-and-fundamental-rights/combatting-discrimination/racism-and-xenophobia/eu-code-conduct-countering-illegal-hate-speech-online_en. Retrieved 2022-08-29. 
  20. ↑ "Code of Practice on Disinformation | Shaping Europe's digital future". September 26, 2018. https://digital-strategy.ec.europa.eu/en/news/code-practice-disinformation. 
  21. ↑ "The 2022 Code of Practice on Disinformation | Shaping Europe's digital future". March 24, 2023. https://digital-strategy.ec.europa.eu/en/policies/code-practice-disinformation. 
  22. ↑ "Procedure File: 2014/2256(INI) | Legislative Observatory | European Parliament". https://oeil.secure.europarl.europa.eu/oeil/popups/ficheprocedure.do?lang=en&reference=2014/2256(INI). 
  23. ↑ COMMUNICATION FROM THE COMMISSION TO THE EUROPEAN PARLIAMENT AND THE COUNCIL Guidance on Article 17 of Directive 2019/790 on Copyright in the Digital Single Market
  24. ↑ "Terrorist content online". https://home-affairs.ec.europa.eu/policies/internal-security/counter-terrorism-and-radicalisation/prevention-radicalisation/terrorist-content-online_en. 
  25. ↑ Reuter, Markus; Rudl, Tomas; Rau, Franziska; Hildebr, Holly. "Why chat control is so dangerous". https://edri.org/our-work/why-chat-control-is-so-dangerous/. 
  26. ↑ Waddell, Kaveh (June 22, 2016). "A Tool to Delete Beheading Videos Before They Even Appear Online". The Atlantic. https://www.theatlantic.com/technology/archive/2016/06/a-tool-to-delete-beheading-videos-before-they-even-appear-online/488105/. 
  27. ↑ "Partnering to Help Curb Spread of Online Terrorist Content | Facebook Newsroom". http://newsroom.fb.com/news/2016/12/partnering-to-help-curb-spread-of-online-terrorist-content/. 
  28. ↑ Abelson, Hal; Anderson, Ross; Bellovin, Steven M.; Benaloh, Josh; Blaze, Matt; Callas, Jon; Diffie, Whitfield; Landau, Susan et al. (2024). "Bugs in our pockets: The risks of client-side scanning". Journal of Cybersecurity 10. doi:10.1093/cybsec/tyad020. 
  29. ↑ Hill, Kashmir (2022-08-21). "A Dad Took Photos of His Naked Toddler for the Doctor. Google Flagged Him as a Criminal.". The New York Times. ISSN 0362-4331. https://www.nytimes.com/2022/08/21/technology/google-surveillance-toddler-photo.html. "A bigger breakthrough came along almost a decade later, in 2018, when Google developed an artificially intelligent tool that could recognize never-before-seen exploitative images of children. [...] When Mark's and Cassio's photos were automatically uploaded from their phones to Google's servers, this technology flagged them." 
  30. ↑ "Google Flagged Parents' Photos of Sick Children as Sexual Abuse". 2022-08-22. https://gizmodo.com/google-csam-photodna-1849440471. "According to Google, those incident reports come from multiple sources, not limited to the automated PhotoDNA tool." 
  31. ↑ Roth, Emma (2022-08-21). "Google AI flagged parents' accounts for potential abuse over nude photos of their sick kids". https://www.theverge.com/2022/8/21/23315513/google-photos-csam-scanning-account-deletion-investigation. "Google has used hash matching with Microsoft's PhotoDNA for scanning uploaded images to detect matches with known CSAM. [...] In 2018, Google announced the launch of its Content Safety API AI toolkit that can “proactively identify never-before-seen CSAM imagery so it can be reviewed and, if confirmed as CSAM, removed and reported as quickly as possible.” It uses the tool for its own services and, along with a video-targeting CSAI Match hash matching solution developed by YouTube engineers, offers it for use by others as well." 
  32. ↑ "StopNCII Announces PhotoDNA Integration and Niantic as New Industry Partner" (in en-gb). https://swgfl.org.uk/magazine/stopncii-announces-photodna-integration-and-niantic-as-new-industry-partner/. 
  33. ↑ "Unfortunate Truths about Child Pornography and the Internet [Feature"]. MUO. December 7, 2012. http://www.makeuseof.com/tag/unfortunate-truths-about-child-pornography-and-the-internet-feature/. 
  34. ↑ Eher, Reinhard; Craig, Leam A.; Miner, Michael H.; Pfäfflin, Friedemann (2011). International Perspectives on the Assessment and Treatment of Sexual Offenders: Theory, Practice and Research. John Wiley & Sons. p. 514. ISBN 978-1119996200. https://books.google.com/books?id=zjXn3zA7HFcC&pg=PA514. 
  35. ↑ Lattanzi-Licht, Marcia; Doka, Kenneth (2004). Living with Grief: Coping with Public Tragedy. Routledge. p. 317. ISBN 1135941513. https://books.google.com/books?id=lr_NNbSPfbAC&q=icmec+children&pg=PA317. 
  36. ↑ Arthur, Charles (July 22, 2013). "Twitter to introduce PhotoDNA system to block child abuse images". The Guardian. https://www.theguardian.com/technology/2013/jul/22/twitter-photodna-child-abuse. 
  37. ↑ Smith, Catharine (May 2, 2011). "Facebook Adopts Microsoft PhotoDNA To Remove Child Pornography". Huffington Post. http://www.huffingtonpost.com/2011/05/20/facebook-photodna-microsoft-child-pornography_n_864695.html. 
  38. ↑ "Adobe & PhotoDNA" (in en). https://www.adobe.com/in/legal/lawenforcementrequests/photodna.html. 
  39. ↑ "Reddit use PhotoDNA to prevent child pornography". March 19, 2020. https://www.redditinc.com/policies/transparency-report-2019. 
  40. ↑ "Discord Transparency Report: July — Dec 2020" (in en). 2021-04-02. https://discord.com/blog/discord-transparency-report-july-dec-2020. 
  41. ↑ "Tech breakthrough announced on the 20th anniversary of IWF’s first child sexual abuse imagery report" (in en-gb). https://www.iwf.org.uk/news-media/news/tech-breakthrough-announced-on-the-20th-anniversary-of-iwf-s-first-child-sexual-abuse-imagery-report/. 
  42. ↑ "Microsoft tip led police to arrest man over child abuse images". The Guardian. August 7, 2014. https://www.theguardian.com/technology/2014/aug/07/microsoft-tip-police-child-abuse-images-paedophile. 
  43. ↑ Salcito, Anthony (December 17, 2009). "Microsoft donates PhotoDNA technology to make the Internet safer for kids". http://blogs.msdn.com/b/microsoftuseducation/archive/2009/12/17/microsoft-donates-photodna-technology-to-make-the-internet-safer-for-kids.aspx. 
  44. ↑ Grossman, Shelby; Pfefferkorn, Riana; Thiel, David; Shah, Sara; DiResta, Renée; Perrino, John; Cryst, Elena; Stamos, Alex et al. (2024). "The Strengths and Weaknesses of the Online Child Safety Ecosystem" (in en). Stanford Internet Observatory, Freeman Spogli Institute for International Studies. doi:10.25740/pr592kc5483. https://purl.stanford.edu/pr592kc5483. 
  45. ↑ "Empowering Industry to Combat Online Child Sexual Abuse with Expanded PhotoDNA Licensing" (in en-US). https://technologycoalition.org/news/the-tech-coalition-empowers-industry-to-combat-online-child-sexual-abuse-with-expanded-photodna-licensing/. 
  46. ↑ "Kindred Tech" (in en-AU). https://kindredtech.org/. 
  • No URL found. Please specify a URL here or add one to Wikidata.



Retrieved from "https://handwiki.org/wiki/index.php?title=PhotoDNA&oldid=4413539"