The latest news about AI and ML Security, Robustness, Safety, Privacy, Trustworthiness, Ethics and Bias by Adversa.ai
Joined September 2018
- Tweets 626
- Following 869
- Followers 2,800
- Likes 1,249
AI Security News retweeted
The best AI research papers on security of the week cover The best AI research papers on security of pupil shapes revealing GAN-generated faces, adversarial examples with makeup application and others.
#SecureAI #TrustedAI #ResponsibleAI
adversa.ai/blog/best-of-adve…
AI Security News retweeted
Release 0-day cyber physical adversal exploit on TikTok O_o? Why not! Check out Adversa AI TikTok account!
#cybersecurity #hacker #AI #facerecognition
tiktok.com/@adversa.ai/video…
Researchers produced a collision in Apple’s new algorithm, but the company says the finding was expected.
#AI #ML #AIsecurity
theverge.com/2021/8/18/22630…
DNNs are now deployed in real-life applications, their vulnerability to adversarial example attacks has raised concerns, especially in the case of safety- and security-sensitive areas.
#AIsecurity #securityofAI #AI #ML
syncedreview.com/2021/08/10/…
AI Security News retweeted
We are happy to announce that the CTO and co-founder of @Adversa_AI Eugene Neelou @eneelou is taking part in @aivillage_dc by @defcon shedding light on the real history of adversarial machine learning. Don’t miss up!
adversa.ai/blog/event/def-co…
861 candidates have successfully completed the “Online Training and Certification Course (OTCC) in Cyber Security, Artificial Intelligence and Machine Learning (AI & ML)” conducted by the DIAT in collaboration with the DRDO.
#AI #ML #AIsecurity
hindustantimes.com/cities/pu…
AI could help to fight cybercrime but it is not a silver bullet, as it can also be exploited by attackers, and here’s what organisations have to keep in mind.
#AI #ML #AIsecurity #securityofAI
techhq.com/2021/07/how-can-a…
Just like any other technology, AI can itself be vulnerable, leading to additional avenues of exploitation and a new class of cyberattacks.
darkreading.com/application-…
Trusted AI news digest: easy-to-abuse AI-based recommendations, deepfakes for phishing campaigns, Facial recognition fraud and others.
#SecureAI #TrustedAI #ResponsibleAI adversa.ai/blog/towards-trus…
AI Security News retweeted
NEW: Face recognition is being used for bank accounts and buildings entry — but it’s hackable. Thousands tried tricking it to apply for unemployment benefits this year, using masks or videos. Some got through. wsj.com/articles/faces-are-t…
The new framework and approach called “UniFad” is used for unified detection of digital and physical face attacks helping to categorize and detect attacks in different face domains applied to the image.
#bestAIresearch2021
arxiv.org/pdf/2104.02156.pdf
[3]
Adversa expresses gratitude for the help in contributing to this report:
@romanyam (University of Louisville)
@janleike (OpenAI)
@VeredShwartz (Allen Institute, University of Washington)
[2]
Adversa expresses gratitude for the help in contributing to this report:
@biggiobattista (University of Cagliari, Pluribus One)
@OliverRochford (Brim Security)
@daviottenheimer (Inrupt, Flyingpenguin)
[1]
Adversa expresses gratitude for the help in contributing to this report:
@adversariel (OpenAI, DefCon AI Village)
Martin Szugat (Deep Learning World)
Avivah Litan (Gartner Research)
Nigel Willson (AwakenAI.org, WeandAI.org)
Now officially, Adversa.ai, an Israeli leader in Secure and Trusted AI, released a report concerning 10 years of Research, Industry, and Government initiatives in Security of AI in the wake of the upcoming EU regulations in AI.
#donttrustAI
adversa.ai/report-secure-and…
The Federal Trade Commission is taking a hard look at AI bias warning companies that selling or using such systems could constitute a violation of federal law.
#AI #ML #AItrustworthiness
protocol.com/ftc-bias-ai
The new “Jekyll” is a framework that takes a biomedical image of a patient as input and translates it to another image indicating an attacker-chosen disease condition.
#AIresearch
arxiv.org/abs/2104.02107
Israel’s National Cyber Directorate has issued a ‘Request for Information’ from providers of KYC checking technologies, to set up authentication capabilities for access to online public services and assist financial regulators.
#AI #ML #AIsecurity
biometricupdate.com/202104/i…
Defending against facial recognition attacks: what they are, why can be dangerous and how to deal with them.
#AI #ML #AIsecurity
thefutureofthings.com/15512-…
The European Union is considering banning the use of artificial intelligence for a number of purposes, including mass surveillance and social credit scores.
#AI #ML #AIsecurity
theverge.com/2021/4/14/22383…