#face-recognition — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #face-recognition, aggregated by home.social.
-
Clearview AI Is Testing an AI Tool That Would Let Cops Unearth Your Life Online
-
DATE: September 7, 2026 at 04:00PM
SOURCE: PSYPOST.ORG** Research quality varies widely from fantastic to small exploratory studies. Please check research methods when conclusions are very important to you. **
-------------------------------------------------TITLE: Researchers find AI spontaneously develops a sense of facial beauty without training
Artificial intelligence models can spontaneously develop an ability to recognize human facial beauty without any prior training or exposure to human preferences. A recent study published in Psychology of Aesthetics, Creativity, and the Arts suggests that this aesthetic sense emerges naturally from the way computer networks process basic visual information. The research hints that human aesthetic preferences might also have an innate origin rooted in our basic sensory wiring.
For centuries, philosophers and scientists have debated whether humans are born with an innate sense of beauty or if aesthetic preferences are learned through cultural exposure. In human infants and nonhuman primates, researchers have observed a natural preference for faces that adults deem attractive. It remains difficult to determine whether this preference is hardwired into the brain before birth or acquired rapidly through early visual experiences.
Deep neural networks, a type of artificial intelligence designed to mimic the human brain’s visual processing pathways, offer a way to test this question. In an untrained state, these networks contain randomized connections and have no prior visual experience or knowledge of human preferences. If an untrained network can detect beauty, it suggests the ability is a byproduct of basic structural encoding rather than a learned behavior.
Researchers Tianxin Shu and Delong Zhang from South China Normal University, along with their colleagues, designed an experiment to explore this phenomenon. They aimed to see if artificial neural units would naturally respond to varying degrees of facial beauty. The research team focused on understanding how simple sensory properties might inherently shape aesthetic perception.
The researchers first generated a dataset of human faces using a specialized artificial intelligence program capable of producing highly realistic images. They manipulated the underlying code of these artificial faces to create different levels of beauty. A small group of human volunteers assessed the images to ensure they still looked like natural human faces.
This process yielded more than 10,000 facial images with distinctly graded aesthetic values. The researchers fed these images into an untrained version of a deep neural network called AlexNet. This network consists of multiple layers of artificial neurons that process visual information, starting with simple edges and moving to whole objects.
Because the network was untrained, its internal weights, which are the mathematical rules determining how neurons interact, were completely random. The researchers analyzed the responses of artificial neurons in the network’s final visual processing layer. They discovered that specific neural units spontaneously began responding selectively to different levels of facial beauty.
These “beauty-selective units” appeared consistently, regardless of the face’s orientation or the specific random initialization of the network. To understand how the network processed this information, the researchers altered the facial images. They applied random scrambling, which disrupted the overall shape of the face but kept local details intact.
They also utilized a different type of scrambling that preserved the general spatial arrangement but obscured fine details. Additionally, they tested images reduced entirely to their simple outlines. The team found that both the specific facial features and the spatial arrangement of those features were necessary for the network to evaluate beauty.
In psychological research, this is known as dual-code theory. This theory suggests that humans rely on both the detailed traits of a face and its broader geometric layout to recognize it. The neural network processed this visual information hierarchically, much like the human visual system.
The network’s earlier layers handled simple geometric features and local details. The later layers then integrated this fragmented data into a complete representation of facial beauty. The research team then tested whether these beauty-selective units could actively compare faces.
They trained a simple sorting algorithm to use the responses from these specific units to judge which of two faces was more attractive. The beauty-selective units successfully differentiated between varying levels of beauty. They performed better than units that simply processed raw image pixels.
The accuracy of these units increased as the difference in beauty between the two faces grew larger. Next, the scientists investigated how the final layer of the network arrived at its beauty preferences. They tracked the activity of units in the earlier layers of the network.
They identified specific neurons that responded to beauty on a sliding scale. The activity of these units steadily increased or steadily decreased as the faces became more attractive. The researchers mapped how these earlier units connected to the final beauty-selective units.
They found that the network constructed its sense of beauty by combining the inputs from these earlier, simpler neurons. Units that preferred highly attractive faces received stronger signals from neurons that increased their activity in response to beauty. Finally, to ensure their results were not unique to one specific network or dataset, the researchers repeated their experiments.
They used a different neural network architecture, known as VGG-16, and again found that beauty-selective units emerged spontaneously. They also tested the network using a database of 200 real human faces spanning different racial backgrounds. The beauty-selective units successfully evaluated the attractiveness of these natural faces.
This outcome proved their ability generalized beyond artificial images. While these findings offer an intriguing look at artificial cognition, a network’s sense of beauty is not identical to human aesthetic appreciation. The study focused on a mathematical abstraction of facial structure rather than the subjective emotional experience humans associate with beauty.
A generalized visual preference for symmetrical or average features does not equate to the rich psychological experience of finding someone beautiful. The artificial faces used in the primary experiments were heavily standardized, minimizing variations in lighting, background, and hairstyle. Real-world aesthetic judgments involve highly variable environments and subjective personal preferences that these models cannot capture.
The current algorithm also primarily evaluated a single baseline standard of beauty. Future research could explore how untrained networks respond to other aesthetic categories, such as natural landscapes, architecture, or fine art. Expanding the criteria for beauty could help scientists determine if this inherent perceptual bias applies to all visual structures or just human faces.
The study, “The Spontaneous Emergence of “A Sense of Beauty ” in Untrained Deep Neural Networks,” was authored by Tianxin Shu, Huawei Xu, Xingxing Chen, Yuxuan Cai, Ming Liu, and Delong Zhang.
-------------------------------------------------
Private, vetted email list for mental health professionals: https://www.clinicians-exchange.org
Unofficial Psychology Today Xitter to toot feed at Psych Today Unofficial Bot @PTUnofficialBot
-------------------------------------------------
#psychology #counseling #socialwork #psychotherapy @psychotherapist @psychotherapists @psychology @socialpsych @socialwork @psychiatry #mentalhealth #psychiatry #healthcare #depression #psychotherapist #AIbeauty #facialaesthetics #untrainedneuralnetworks #deeplearning #AlexNet #VGG16 #beautyselection #neuroscienceofaesthetics #dualcodetheory #facerecognition
-
#FaceRecognition Is Becoming the Norm for #Dating Apps
To keep #scammers at bay, dating apps are embracing #biometric scanning tools and “verified human” badges. These moves are sparking concerns about #surveillance and #privacy.
https://www.wired.com/story/face-recognition-is-becoming-the-norm-for-dating-apps/
-
ICYMI: Explaining facial age estimation: Facial age estimation predicts age from a face image without identifying the person: the mechanics, the NIST accuracy record, and the open regulatory dispute. https://ppc.land/facial-age-estimation/ #FacialAgeEstimation #AgePrediction #FaceRecognition #NIST #AI
-
Face Recognition Is Becoming the Norm for Dating Apps
-
@eff and Civil Society Groups Call on #Nottinghamshire #Police to Halt Live #FaceRecognition
#privacy #biometrics -
Reverse-Lookup Service Exposed Millions of Photos of People’s Faces
-
When #AI face recognition meets GDPR, absurdity ensues:
"When someone is identified, a 2D image flashes up on the store's iPhone. But, and this is wild, for GDPR reasons, it disappears after a few seconds once staff have seen it.
So even the store manager has no idea whether staff saw an image of someone who looked a bit like me and cocked up, or whether it was actually me and the AI cocked up."https://cybernews.com/ai-news/ai-facial-recognition-wrongly-ejects-comedy-worker-donates-payout/
-
New York City Lawmakers Push to ‘Ban the Scan’ at MSG
-
Why Biometric Identification Is Becoming A Core Feature of Crypto Security
Biometric technologies use physical features to enable identity verification, such as voice, fingerprints, or facial features. These methods help keep criminals from stealing data to carry out illicit transactions, which is why crypto exchanges and individual users are increasingly integrating them. Exchanges utilize technologies such as electronic Know Your Customer (eKYC) and liveness detection to create a multilayered security system…….Continue reading…. By Samuel […] -
Adversarial Clothing Designed to Fool Facial Recognition Systems
There are many companies manufacturing adversarial clothing designed to confuse ... https://www.schneier.com/blog/archives/2026/08/adversarial-clothing-designed-to-fool-facial-recognition-systems.html
#facerecognition #securitytheater #Uncategorized #surveillance
-
Hey, Illinois, go sign up! Join the class action suit (again) against Apple, get $5K!
~Users did not consent to having their faces scanned and matched by Apple
~Other states have won similar fights
~6.5 million can join
~If you had pictures in your Apple Photos after 2016, you may qualify!
#fascist #USA #surveillance #state #BigBrother #flock #cameras #Apple #FaceRecognition #scanning #privacy #laws #folders #photo #app #library
https://www.pcmag.com/news/apple-to-face-325-billion-lawsuit-over-facial-recognition-in-photos-app -
Looks like more money might be heading back to your wallet.
Judge greenlights multi-billion-dollar class action over Apple Photos’ face recognition - 9to5Mac
-
Facial Recognition at Madison Square Garden
Last month, the story broke (alternate link) that Madison Square Garden uses facial recogn... https://www.schneier.com/blog/archives/2026/07/facial-recognition-at-madison-square-garden.html
#facerecognition #Uncategorized #surveillance #stalking #privacy
-
did somebody test the #smartdepart app yet? probably soon mandatory everywhere in germany, wenn das so weitergeht...
https://ber.berlin-airport.de/de/ausgewaehlt/ber-biometrics.html
https://www.youtube.com/shorts/FcNvQge8Omc
#berlin #airport #ber #biometrics #surveillance #facerecognition
-
States Want ICE Agents to Show Their Faces. The Trump Administration Is Blocking Them
-
Amazing and fun workshop using make-up to fool face recognition software.
Thanks @vecna and N.I.N.A. for bringing it to Genova!
You can play yourself with Ghostmaxxing
#Genova #G8 #G8Genova #surveillance #surveillanceState #faceRecognition #surveillance #stateSurveillance #makeup
-
From #LouisRossmann, for anyone concerned with #surveillance and #faceRecognition:
The image recognition component of #Google #Gemini insists that multiple contemporary photos of American repair technician and #rightToRepair advocate Steven Rhine are definitely photos of an (obviously long-dead) #Sep11 hijacker.
As is said in the video, “what the #flock!”
(Yes, it’s a different image matching program returning a false positive, but Flock warrants becoming a curse.)
-
Meta's NameTag Face Recognition Technology: Existence and Legal Implications
📰 Original title: Here’s the Truth About Whether Meta’s NameTag Face Recognition Tech ‘Exists’
🤖 IA: It's clickbait ⚠️
👥 Users: It's clickbait ⚠️View full AI summary https://en.killbait.com/meta-s-nametag-face-recognition-technology-existence-and-legal-implications.html?utm_source=mastodon_world&utm_medium=social&utm_campaign=killbait.mastodon_world
-
Meta's NameTag Face Recognition Technology: Existence and Legal Implications
📰 Original title: Here’s the Truth About Whether Meta’s NameTag Face Recognition Tech ‘Exists’
🤖 IA: It's clickbait ⚠️
👥 Users: It's clickbait ⚠️View full AI summary https://en.killbait.com/meta-s-nametag-face-recognition-technology-existence-and-legal-implications.html?utm_source=mastodon_social&utm_medium=social&utm_campaign=killbait.mastodon_social
-
Here’s the Truth About Whether Meta’s #NameTag #FaceRecognition Tech ‘Exists’
Since WIRED reported on Meta’s NameTag face recognition system, company executives have made confusing and conflicting remarks about its very existence.
#meta #privacy #security #biometricshttps://www.wired.com/story/heres-the-truth-about-whether-metas-nametag-face-recognition-exists/
-
Here’s the Truth About Whether Meta’s NameTag Face Recognition Tech ‘Exists’
-
#Chameleon (P3-Mask), generates personalized privacy protection masks designed to thwart unauthorized facial recognition while preserving image quality. It uses cross-image optimization, perceptibility tuning, and a robustness-focused ensemble approach. Try it on GitHub: https://github.com/git-disl/Chameleon 🤖🛡️ #ComputerVision #PrivacyProtection #FaceRecognition #P3Mask
-
@eff , #TEDIC and #CEJIL Challenge #Secrecy in the Use of #FaceRecognition in #Paraguay
Seeking #transparency and #accountability in Paraguay’s use of facial recognition, EFF, the Association of Technology, Education, Development, Research, Communication (TEDIC), and the Centre for Justice and International Law (CEJIL) filed a complaint with the Inter-American Commission on #HumanRights against the state for arbitrarily denying access to information about its implementation and use of the technology as a tool for mass #surveillance that erodes people’s #privacy #rights.
-
EFF, TEDIC and CEJIL Challenge Secrecy in the Use of Face Recognition in Paraguay
Published in Electronic Frontier Foundation (EFF)
Seeking transparency and accountability in Paraguay’s use of facial recognition, EFF, the Association of Technology, Education, Development, Research, Communication (
#Blog #Democracy #InformationAccess #DigitalRights #FaceRecognition #Privacy
-
🚨 Breaking news: Your face is now the hottest commodity since avocado toast! 🥑😱 Big Brother's latest scheme involves "protecting" children by turning our selfies into ID cards. Just remember, no mask can save your mug from their Orwellian obsession with your age. 😂📸
https://nevergivethemyourface.com/ #BreakingNews #FaceRecognition #BigBrother #Privacy #Concerns #OrwellianSociety #IDCards #HackerNews #ngated -
Meta Deletes Face-Recognition System From Its Smart Glasses App After WIRED Report
https://iphone.socnaurom.workers.dev/vi/post/1047
Meta Deletes Face-Recognition System From Its Smart Glasses App After WIRED Report Read Full Article #FaceRecognition #TechNews #MetaNews Meta Deletes Face-Recognition System From Its Smart Glasses App After WIRED Report Read Full Article #FaceRecognition #TechNews #MetaNews...
via Tin tức AI Toàn Cầu | TechOffice AI News https://iphone.socnaurom.workers.dev/vi/
June 9, 2026 at 06:15PM