Tag Archive for: Accuracy

A New Attack Reveals Everything You Type With 95 Percent Accuracy


Of course, generative AI tools are the talk of the security industry this year. And Microsoft is no exception. In fact, since 2018, the company has had an AI red team that attacks AI tools to find vulnerabilities and help prevent them from behaving badly.

Outside of Black Hat and Defcon coverage, we detailed the ins and outs of the data privacy that HIPPA provides people in the US, and explained how to use Google’s new “Results About You” tool to get your personal information removed from search results.

But that’s not all. Each week, we round up the security news that we didn’t cover in depth ourselves. Click on the headlines to read the full stories. And stay safe out there.

Your keyboard may be exposing your secrets without you even knowing it. Researchers in the UK developed a deep-learning algorithm that can figure out what a person is typing just by listening to keystrokes. In a best-case scenario (for an attacker, that is), the algorithm is 95 percent accurate. The researchers even tested it over Zoom and found it performed with 93 percent accuracy.

Now, if you’re thinking the researchers tested the attack on the noisiest mechanical keyboard they could find, you’d be wrong. They performed their tests on a MacBook Pro. And the attack doesn’t even require fancy recording equipment—a phone’s microphone works just fine. Someone who successfully carries out the attack could use it to learn a target’s passwords or snoop on their conversations. These kinds of acoustic attacks aren’t new, but this research shows they’re getting frighteningly accurate and easier to pull off in the wild.

A series of data breaches rocked the United Kingdom this week. On August 8, the Electoral Commission, the independent body responsible for overseeing elections and regulating political finances, revealed a cyberattack had exposed the data of 40 million voters to hackers. The organization has been unable to determine whether data was taken; however, it says that full names, emails, phone numbers, home addresses, and data provided during contact with the body could be impacted. “The attack has not had an impact on the electoral process,” the commission said. (Elections are run…

Source…

“China Is Watching” – With AI-Powered Satellites & Thousands Of Cameras, Can Beijing Strike Key US, UK Targets With Pinpoint Accuracy?


Last month, Fraser Sampson, Britain’s Commissioner for Biometrics and Surveillance Cameras, wrote to Cabinet Minister Michael Gove to convey his concerns about the dominance of Chinese video surveillance equipment in Britain.

He said he had “become increasingly concerned at the security risks presented by some state-controlled surveillance systems covering our public spaces.”

Two Chinese companies, Hikvision and Dahua, have grabbed a huge share of Britain’s CCTV market. While both, Hikvision, which has revenues of $9.3 billion, and Dahua, whose revenues are $3.7 billion, are private companies but they have major shareholders with ties to the Chinese Communist Party (CCP).

Also, Hikvision is known to be controlled by China Electronics Technology Group Ltd. (CETC), one of the major Chinese military-industrial groups, and is China’s largest electronics defense contractor. Under the Chinese Communist Party (CCP) regime, all military-industrial groups have to obey the orders of the regime.

Tiangong_Space_Station-China
File Image: Tiangong Space Station – China

Thousands Of Cameras In Britain

The UK-based campaign group Big Brother Watch sent 4,500 freedom of information (FoI) requests to public bodies asking whether they had Hikvision or Dahua cameras employed on their premises.

Of the 1,300 who responded, 800 confirmed that they did, including nearly three-quarters of councils, 60% of schools, half of NHS trusts and universities, and nearly a third of police forces.

Moreover, Big Brother Watch found that there are 164,000 Hikvision cameras and 14,000 Dahua cameras in public spaces apart from the government bodies.

Reports suggest that many of these cameras have advanced features such as microphones, the capacity for facial and gender recognition, and distinguishing between people of different racial groups.

hypersonic
File Image: Hypersonic Missile

Some cameras can also analyze behavior, such as detecting if a fight might be breaking out. Others can even judge moods, track via heat-sensing, and learn behavior patterns to highlight any unusual activity.

Backdoors Discovered In Chinese-Made Cameras

Serious security flaws have been detected in the past in cameras produced by both of these companies, which could…

Source…

Computer vision can help spot cyber threats with startling accuracy


This article is part of our reviews of AI research papers, a series of posts that explore the latest findings in artificial intelligence.

The last decade’s growing interest in deep learning was triggered by the proven capacity of neural networks in computer vision tasks. If you train a neural network with enough labeled photos of cats and dogs, it will be able to find recurring patterns in each category and classify unseen images with decent accuracy.

What else can you do with an image classifier?

In 2019, a group of cybersecurity researchers wondered if they could treat security threat detection as an image classification problem. Their intuition proved to be well-placed, and they were able to create a machine learning model that could detect malware based on images created from the content of application files. A year later, the same technique was used to develop a machine learning system that detects phishing websites.

The combination of binary visualization and machine learning is a powerful technique that can provide new solutions to old problems. It is showing promise in cybersecurity, but it could also be applied to other domains.

Detecting malware with deep learning

The traditional way to detect malware is to search files for known signatures of malicious payloads. Malware detectors maintain a database of virus definitions which include opcode sequences or code snippets, and they search new files for the presence of these signatures. Unfortunately, malware developers can easily circumvent such detection methods using different techniques such as obfuscating their code or using polymorphism techniques to mutate their code at runtime.

Dynamic analysis tools try to detect malicious behavior during runtime, but they are slow and require the setup of a sandbox environment to test suspicious programs.

In recent years, researchers have also tried a range of machine learning techniques to detect malware. These ML models have managed to make progress on some of the challenges of malware detection, including code obfuscation. But they present new challenges, including the need to learn too many features and a virtual environment to analyze the target samples.

Binary visualization can…

Source…

Tor traffic from individual Android apps detected with 97 percent accuracy – ZDNet

Tor traffic from individual Android apps detected with 97 percent accuracy  ZDNet

New machine learning algorithm can detect when you’re using a specific app, such as YouTube, Instagram, Spotify, others.

“android security news” – read more