Intel creates worlds first real-time fake video detector called FakeCatcher

November 22, 2022 at 12:22PM

As more users are using digital platforms to get information and stay updated about the developments around the world, the threat of misinformation reaching a larger audience is also increasing. Although several tools can source data that has surfaced on the internet, it is difficult to check whether a particular video is fake. To solve the problem, the California-based chip manufacturer has developed a tool called FakeCatcher. It uses AI to detect fake videos with 96% accuracy.

What is Intel’s FakeCatcher and how does it work?

Per Intel, FakeCatcher "is the world's first real-time deepfake detector that returns results in milliseconds." The fake video detector is designed by Demir in collaboration with Umur Ciftci from the State University of New York at Binghamton. 

The technology uses Intel's software and hardware and functions as a web-based tool. The tool runs face and landmark detection algorithms for checking a video and can analyse up to 72 different streams simultaneously on third Gen Intel Xeon processors. 

The other deep learning-based fake video detection tools look at a video's raw data to find a sign of inauthenticity. Whereas Intel's FakeCatcher looks for clues in real videos. In the official press release, the company mentions that the tool detects subtle blood flow in the pixels occupied by human skin in the video, interprets them in spatiotemporal maps and detects whether a video is fake or real. 

The fake video detection tool can be very useful

The press release adds that "Deception due to deepfakes can cause harm and result in negative consequences, like diminished trust in the media." FakeCatcher can help restore trust by enabling individuals, organisations and other entities to distinguish between real and fake content.

While social media platforms can use the tool to detect whether a user is uploading a fake video and keeping the footage from circulating on the forum, it can also be helpful for fact-checking organisations. Additionally, media houses can use the tool to avoid amplifying manipulated videos.

For more technology news, product reviews, sci-tech features and updates, keep reading Digit.in or head to our Google News page. 

from Security Software News https://ift.tt/lChiF72

Indias cyber agency now warns about bugs in Mozilla Firefox browser

August 31, 2022 at 01:09PM

New Delhi, Aug 31 (IANS) After warning users about bugs in Google Chrome for desktop, the Indian Computer Emergency Response Team (CERT-In) has now cautioned against multiple vulnerabilities in Mozilla Firefox products that can let hackers compromise devices and systems.

The bugs in Mozilla Firefox browser could allow a remote attacker to bypass security restrictions, execute arbitrary code and cause denial of service attack on the targeted system, CERT-In said in its latest advisory.

"These vulnerabilities exist in Mozilla Firefox due to abuse of XSLT error handling, cross-origin iframe referencing an XSLT document... that results in a use-after-free error and memory safety bugs within the browser engine," explained the cyber agency.

A remote attacker could exploit these vulnerabilities by convincing a victim to open a specially-crafted web request.

CERT-In, which comes under the IT Ministry, advised users to update to the latest Mozilla Firefox versions.

CERT-In also found a vulnerability in open source coding platform Drupal which could allow an attacker to bypass security restrictions on the targeted system.

"Successful exploitation of this vulnerability could allow an attacker to bypass security restrictions (leak valid payment details and accept invalid payment details) on the targeted system," it warned.

Last week, the cyber agency had warned users about multiple vulnerabilities in Google Chrome for desktop that could let threat actors gain access to their computers.

(Except for the headline, the rest of this IANS article is un-edited)

For more technology news, product reviews, sci-tech features and updates, keep reading Digit.in

from Security Software News https://ift.tt/iAongpI

Ransomware attacks have reportedly surged to over 12 million a month: Who are all affected

August 28, 2022 at 09:31PM

The volume of ransomware threats detected spiked to more than 1.2 million per month between the January-June period, a report has revealed. In the past 12 months, cyber-security researchers at Barracuda Networks identified and analysed 106 highly publicised ransomware attacks. They found the dominant targets are still five key industries: education, municipalities, healthcare, infrastructure and finance. Researchers also saw a spike in the number of service providers that have been hit with a ransomware attack. "Ransomware attackers remain defiant and continue to operate their business with extended extortion attempts. As ransomware and other cyberthreats continue to evolve, the need for adequate security solutions has never been greater," said Parag Khurana, Country Manager, Barracuda Networks India. While attacks on municipalities increased only slightly, the analysis showed that ransomware attacks on educational institutions more than doubled, and attacks on the healthcare and financial verticals tripled. Many cybercriminals target small businesses to gain access to larger organisations. As a result, it is essential for security providers to create products that are easy to use and implement, regardless of a company's size, said the report. Most ransomware attacks don't make headlines, though. Many victims choose not to disclose when they get hit, and the attacks are often sophisticated and extremely hard to handle for small businesses, said the report.

(Except for the headline and cover image, the rest of this IANS article is un-edited)

For more technology news, product reviews, sci-tech features and updates, keep reading Digit.in.

from Security Software News https://ift.tt/byOnpAB

More than 50 of Bitcoin transactions are reportedly fake: Find the details here

August 28, 2022 at 09:17PM

More than half of all Bitcoin daily trading volumes are reportedly fake, as the world's largest cryptocurrency nosedives again amid global economic turmoil. According to a Forbes analysis of 157 crypto exchanges and trading platforms, more than half (51 per cent) of all reported trading volume is likely to be fake or non-economic. "We estimate the global daily bitcoin volume for the industry was $128 billion on June 14. That is 51 per cent less than the $262 billion one would get by taking the sum of self-reported volume from multiple sources," the report mentioned. Bitcoin represents 40 per cent of the $1 trillion global crypto market. According to the report, there is no genuine method of calculating bitcoin daily volume, "even among the industry's most reputable research firms". "For example, CoinMarketCap puts the latest 24-hour trading of bitcoin at $32 billion, CoinGecko at $27 billion, Nomics at $57 billion and Messari at $5 billion," the report revealed. In terms of how much Bitcoin activity takes place, 21 crypto exchanges generate $1 billion or more in daily trading activity, while the next 33 exchanges had volume between $200 million and $999 million. Binance is the leader with a 27 per cent market share, followed by FTX. Chicago-based CME Group is the market leader in bitcoin futures trading. After showing some stabilisation in the last few weeks, world's largest cryptocurrency Bitcoin has plunged once again below $20,000 after Federal Reserve Chair Jerome Powell's keynote address. Bitcoin prices had stabilised around the $23,000 to $24,000 level after plunging below $20,000 in June.

(Except for the headline and cover image, the rest of this IANS article is un-edited)

For more technology news, product reviews, sci-tech features and updates, keep reading Digit.in.

 

from Security Software News https://ift.tt/bviNpaY

Google shares tips on improving the Search experience while using quotes

August 07, 2022 at 06:52PM

Today, Google has revealed some improvement for “quoted” searches. For example, if you type any phrase in the search bar, the snippets will display the accurate location of the phrase. By clicking the link and approaching the content, it will be easier to locate them. Moreover, google stated that they received feedback from users signalling that users of quoted searches want to view the exact location of the said phrase on the website.

Here’s how quoted search works

Google has shared some tips and tricks on how quoted searching works along with some limitations of the feature. 

Searches for quotes might match different results or content that is not clearly seen on the page: Quoted searches sometimes come up or match with the content that is not visible on the page and makes it look like the content is not there even when it is there. Moreover, web pages may also change since the time you last visited them. Google cached copies may display wherein referred to information appeared on the internet page. Use Developer Tools: Some people use the browser find command to find phrases. If it doesn't work, they can also use a developer tool in which users can search for all rendered text which includes drop-down menu text.  Quoted text may appear only in title links: Quoted content will not appear in snippets even if they are only in title links or URLs. It doesn't bold the URL matches and title links. Sometime punctuations seen as spaces: According to Google, punctuation is seen as spaces which affect quoted searches. So they will match text where punctuation like hyphens or commas breakup the words as the letter patterns are the same i.e., don’t , doesn't ; don’t/doesn’t; don’t-doesn’t. Quoted content is bolded for web page snippets on desktop: The new bolding of quoted text works only on the desktop which will not appear in recipe or video boxes and in picture or news searches. The boxes and modes contain quoted phrases and mobile results are not bolded.  Quoted searches don’t work for local results: For local results within maps, the quote function does not function.

Yonghao Jin, software engineer, Search, Google, while speaking on the improvement of quoted searches, said, By default, our system is designed to look for both the right text and phrase entered related terms and concepts, which is often useful. 

He added, if you use quotes you might miss helpful content that uses closely related words. Of course, there are those times when the exact words being on a web page makes all the difference. For those situations, quoted searches remain available and are now even better. 

For more technology news, product reviews, sci-tech features and updates, keep reading Digit.in

from Security Software News https://ift.tt/vQs1qyM

Google Spots A New Spyware On Android And iOS: Should You Be Worried

June 24, 2022 at 07:00AM

Google’s Project Zero and Threat Analysis Group (TAG) has come forward with its findings on the activities of an Italian spyware maker named RCS Labs. This is not as big in scale or scope as Israeli NSO Group and its proprietary Pegasus spyware. Nonetheless, it has reportedly been around for quite a few years and has been used on people in Italy, Kazakhstan, and Syria. Even if your country's name isn’t on the list, know that TAG is currently tracking more than 30 spyware vendors that have grown into a full-blown ecosystem and lends their services to world governments. So, let’s understand how these things work.

How Do RCS Labs’ Android And iOS Spyware Work?

The spyware will be masked as a fake My Vodafone app that is pushed to the users through an SMS link and they are tricked into installing the app. Well, to convince them, the attackers have sometimes got the ISPs to disconnect the mobile data first and then ask them to install the particular My Vodafone app to restore the services.

The app would seem legit and the sideloading works because it was signed in through Apple's Enterprise Developer Program. Apple has however revoked all certificates and accounts related to this now.

Talking about sideloading, Apple said, “Enterprise certificates are meant only for internal use by a company, and are not intended for general app distribution, as they can be used to circumvent App Store and iOS protections. Despite the program’s tight controls and limited scale, bad actors have found unauthorized ways of accessing it, for instance by purchasing enterprise certificates on the black market.”

Apple has also patched the exploits that were used by the bad actors to sneak into the victim’s iPhones. 

According to Project Zero member Ian Beer, the exploits were successful in the first place, because of the new  “system-on-a-chip” and “coprocessors” used in the recent iPhones, something which is used by Android phones too.

Meanwhile, TAG member Benoit Sevens remarked, “The commercial surveillance industry benefits from and reuses research from the jailbreaking community. In this case, three out of six of the exploits are from public jailbreak exploits. We also see other surveillance vendors reusing techniques and infection vectors initially used and discovered by cyber crime groups. And like other attackers, surveillance vendors are not only using sophisticated exploits but are using social engineering attacks to lure their victims in.”

Another TAG employee Clement Lecigne told WIRED that “These vendors are enabling the proliferation of dangerous hacking tools, arming governments that would not be able to develop these capabilities in-house. But there is little or no transparency into this industry, that's why it's critical to share information about these vendors and their capabilities.”

We agree and appreciate Google and other parties involved in discovering such vulnerabilities. Now if you own an iPhone or for that matter any computing device, you are advised to keep their software up to date.

As for other news, reviews, feature stories, buying guides, and everything else tech-related, keep reading Digit.in.

from Security Software News https://ift.tt/dvmQgPh