#FHIBE
A paper in Nature presents a database of more than 10,000 human images to evaluate biases in artificial intelligence models for human-centric computer vision, called the Fair Human-Centric Image Benchmark (FHIBE) and was developed by Sony AI. go.nature.com/4qP0MAK 🧪
November 6, 2025 at 2:13 PM
🔍 Behind the paper: 'FHIBE: A New Standard for Ethical AI Evaluation.' Introduces the world’s first consent-driven, globally diverse dataset for fairness benchmarking in computer vision—centering human dignity, privacy, and bias evaluation. 🔗 Read more here: bit.ly/4ahIAZY. #STS #PhilSci
January 30, 2026 at 4:00 PM
All AI models are constrained by the quality of their data. A recent @nature.com paper introduces the Fair Human-Centric Image Benchmark (FHIBE), a diverse dataset implementing best practices for consent, privacy, compensation, safety and utility. 🧪 www.nature.com/articles/s41...
Fair human-centric image dataset for ethical AI benchmarking - Nature
The Fair Human-Centric Image Benchmark (FHIBE, pronounced ‘Feebee’)—an image dataset that implements best practices for consent, privacy, compensation, safety, diversity and utility—can be used respon...
www.nature.com
December 8, 2025 at 7:09 PM
Computer vision, on which many AI applications rely on, is based on non-consensual datasets, compromising fairness and accuracy in AI. Now, an ethically curated, diverse human-image data set, sets a new standard for responsible data curation and trustworthy AI. 🧪 www.nature.com/articles/s41...
Fair human-centric image dataset for ethical AI benchmarking - Nature
The Fair Human-Centric Image Benchmark (FHIBE, pronounced ‘Feebee’)—an image dataset that implements best practices for consent, privacy, compensation, safety, diversity and utility—can be used respon...
www.nature.com
November 12, 2025 at 2:58 PM
Sony FHIBE benchmark exposes gaps in AI fairness today

Sony FHIBE benchmark debuts a consent-based dataset to audit bias across computer vision, revealing stereotype risks and pronoun gaps in AI.

https://www.aistory.news/ai-ethics-and-regulation/sony-fhibe-benchmark-exposes-gaps-in-ai-fairness-...
Sony FHIBE benchmark exposes gaps in AI fairness today
Sony FHIBE benchmark debuts a consent-based dataset to audit bias across computer vision, revealing stereotype risks and pronoun gaps in AI.
www.aistory.news
November 5, 2025 at 8:05 PM
Alice Xiang, Global Head of AI Governance at Sony Group Corporation, and Lead Research Scientist at Sony AI, on FHIBE: Global diversity. True consent. Scientific rigor.
Watch A Fair Reflection + explore fairnessbenchmark.ai.sony
#FHIBE #FairAI #AIFairness #EthicalAI #SonyAI
December 4, 2025 at 10:21 PM
🚨 Big news: FHIBE is here.
Sony AI’s Fair Human-Centric Image Benchmark, published in Nature, is the first globally diverse, consent-driven dataset to benchmark AI fairness.

Explore the research + dataset: bit.ly/3JKaqnm
#FHIBE #AIethics #SonyAI
November 19, 2025 at 8:07 PM
Important image dataset #FHIBE obtained with informed consent, compensation, safety measures, bias reduction and collection cost transparency, demonstrating that “implementing core ethical considerations in practice is possible”.

#ethicaldatacollection
Fair human-centric image dataset for ethical AI benchmarking - Nature
The Fair Human-Centric Image Benchmark (FHIBE, pronounced ‘Feebee’)—an image dataset that implements best practices for consent, privacy, compensation, safety, diversity and utility—can be used respon...
www.nature.com
November 6, 2025 at 3:31 PM
🚨 Say hello to FHIBE! 🌍 Sony's new dataset is tackling bias and fairness in AI with consent-based data collection—paving the way for responsible AI. What are your thoughts? 🤖✨ #EthicalAI #FairnessInAI #SonyAI LINK
November 6, 2025 at 8:18 AM
A database of more than 10,000 human images to evaluate biases in artificial intelligence (AI) models for human-centric computer vision is presented in Nature this week: spklr.io/63322BzLxu
Fair human-centric image dataset for ethical AI benchmarking - Nature
The Fair Human-Centric Image Benchmark (FHIBE, pronounced ‘Feebee’)—an image dataset that implements best practices for consent, privacy, compensation, safety, diversity and utility—can be used responsibly as a fairness evaluation dataset for many human-centric computer vision applications.
spklr.io
November 5, 2025 at 8:21 PM
“FHIBE also determined that today's AI models reinforced stereotypes when prompted with neutral questions about a subject's occupation. The tested models were particularly skewed ‘against specific pronoun and ancestry groups,’ describing subjects as sex workers, drug dealers or thieves.”
November 5, 2025 at 5:34 PM
The Fair Human-Centric Image Benchmark (FHIBE, pronounced ‘Feebee’)—an image dataset that implements best practices for consent, privacy, compensation, safety, diversity and utility—can be used responsibly as a fairness evaluation dataset for many human-centric computer vision applications.
Fair human-centric image dataset for ethical AI benchmarking - Nature
www.nature.com
November 6, 2025 at 1:07 AM
Sony släpper verktyg för att mäta bias i AI-modeller

Sony har lanserat ett nytt mätverktyg, FHIBE (Fair Human-Centric Image Benchmark), för att bedöma hur rättvisa och opartiska dagens AI-modeller är. Förhoppningen är att FHIBE kan användas för att identifiera och adressera bias i AI-system.…
Sony släpper verktyg för att mäta bias i AI-modeller
Sony har lanserat ett nytt mätverktyg, FHIBE (Fair Human-Centric Image Benchmark), för att bedöma hur rättvisa och opartiska dagens AI-modeller är. Förhoppningen är att FHIBE kan användas för att identifiera och adressera bias i AI-system. Datainsamlingen består av 10 000 bilder på cirka 2 000 betalda deltagare från över 80 länder, som var och en har samtyckt och kan begära borttagning av sina bilder när som helst.
folketstidning.wordpress.com
November 6, 2025 at 6:45 PM
へぇ、さすがソニー。被写体の同意を得て倫理観のチェックに使える画像データセット、学術誌にも評価される内容、とな。解説でCSLに言及があるのは北野さんも絡んでいる動きなのかな、流石 // AIの学びに偏りはないか ソニーG、人の画像見せ倫理観判定 - 日本経済新聞 https://www.nikkei.com/article/DGXZQOUC2798H0X21C25A1000000/
AIの学びに偏りはないか ソニーG、人の画像見せ倫理観判定 - 日本経済新聞
ソニーグループの人工知能(AI)研究開発部門であるソニーAIが、人間の画像をAIが分析した結果にバイアス(偏り)がないかどうかを評価するデータセット「Fair Human-Centric Image Benchmark(FHIBE、フィービー)」を公開した。AIモデルの学習ではなく、評価のために使うといった一定の条件下で、AI開発企業などが無料で使える。ソニーGのAIガバナンス責任者、アリス・シャン氏に狙いを聞いた。(聞き手はシリコンバレー=中藤玲)
www.nikkei.com
November 29, 2025 at 5:49 AM
Sony AI hat mit dem Fair Human-Centric Image Benchmark (FHIBE) einen neuen Standard zur Messung von Bias in Computer-Vision-Modellen veröffentlicht. Der Datensatz umfasst über 10.000 Bilder aus 81 Ländern, alle mit Zustimmung erhoben – ein wichtiger Schritt zu mehr Transparenz und Fairness im […]
Original post on det.social
det.social
November 7, 2025 at 6:42 AM
Sony onthult schokkende bias in AI-modellen
Sony onthult schokkende bias in AI-modellen
Sony's nieuwe testdataset FHIBE toont aan dat alle geteste AI-modellen significante vooroordelen vertonen bij het behandelen van mensen. De bevindingen zijn ronduit verontrustend.
www.vandaagai.nl
November 5, 2025 at 4:55 PM
Sony has a new benchmark for ethical AI

Sony AI released a dataset that tests the fairness and bias of AI models. It's called the Fair Human-Centric Image Benchmark (FHIBE, pronounced like "Phoebe"). The company describes it as the "first publicly available, globally diverse, consent-based human…
Sony has a new benchmark for ethical AI
Sony AI released a dataset that tests the fairness and bias of AI models. It's called the Fair Human-Centric Image Benchmark (FHIBE, pronounced like "Phoebe"). The company describes it as the "first publicly available, globally diverse, consent-based human image dataset for evaluating bias across a wide variety of computer vision tasks." In other words, it tests the degree to which today's AI models treat people fairly.
toolcome.com
November 5, 2025 at 7:20 PM