IEAI, August 2022
What is an audit for artificial intelligence systems? An audit is a procedure that allows you to get a complete overview of the current and past performance of an AI system. Why is it needed? Audit for AI systems is essential to proactively eliminate, mitigate or manage potential risks. How to carry out an audit for AI systems? The AI audit can be carried out both within the organization developing AI technology, and by an external company.
In recent years, research has been carried out on the safe use of AI, and the number of incidents is growing inexorably. As a result, industry, academia, non-profits, and government agencies have published numerous guidelines for developing secure AI. For example, last year, in order to ensure the development of secure AI systems in line with the European Union’s human-centered values, the EU Law on Artificial Intelligence was issued as a proposal for the regulation of AI systems. The creation of a new Defense Center for AI Research (DCAR) will be part of Defense Artificial Intelligence Center (DAIC) in the UK. In addition, major tech companies such as Google and Microsoft have implemented AI Red Teams to uncover weaknesses, biases, and security flaws of their machine learning systems.
This paper is a brief study on the topic of AI auditing and the need for interdisciplinary discussion of this topic. AI system auditing has been widely recognized as an opportunity to harness the potential of AI models, and to detect and alleviate problem patterns and the consequences of their deployment in sensitive contexts.
CNET, August 8, 2022
According to VMware’s annual Global Incident Response Threat Report, there has been a 13% increase in deepfakes and cyberattacks that give the false impression of someone saying or doing something. This happens because the technology threat is shifting from hypothetical to real harm, a new report says.
Deepfake technology uses artificial intelligence to make it look like a person is doing or saying something they are not actually doing. It has become popular in 2019. But while early deepfakes were mostly easy to detect, the technology has now evolved to be much more forceful. For example, in March 2022, a video appeared on social networks containing an appeal Ukrainian President Volodymyr Zelenskyy to his soldiers with an order to surrender to Russian troops. Zelenskyy was quick to denounce it, but showed the potential harm deepfakes could cause.
Recently, the FBI notified that scammers were using deepfakes to interview for remote jobs in areas like information technology, programming, and other software-related roles, as well as trying to pass on personal information stolen to pass a background check.
The Machine Learning Security Evasion Competition 2022 (MLSEC 2022) is a collaboration between Zoltan Balazs (Head of Vulnerability Research Lab at CUJO AI), Hyrum Anderson (Distinguished Engineer at Robust Intelligence), and Eugene Neelou (Co-Founder and CTO at Adversa AI) to allow researchers to exercise their attacker muscles against ML security models in a unique real-world setting.
The aim of the contest is to raise awareness of the various ways in which threat actors can evade ML systems. “Watching the evasions evolve during the competition over the years is very exciting for us,” admits Balazs. “We’re really excited about creating this opportunity to get researchers and cybersecurity professionals together to exercise their skills in a unique real-world setting.”
This year, the competition is organized around two separate tracks of the attacker challenge:
What are you waiting for?! It’s your chance to win up to $9000 by participating in this event!
Stay up to date with what is happening! Get a first look at news, noteworthy research and worst attacks on AI delivered right in your inbox.
Four months ago, Adversa presented the world’s first Fool AI Art Exhibition to the public. We tricked the AI-driven face recognition system 100 times, released it as the world-first piece ...
