Disinformation, right-wing influencers and the US election.
Sextortion exploits Street View. Criminal use of AI accelerates. And Telegram starts moderating private chats.
Election games
As the US presidential election campaign enters its final weeks, the extent of foreign meddling is clear to see. This week, the US Justice Department carried out a widespread crackdown on Russia’s disinformation operations, while revealing the extent to which well-known right-wing influencers have been part of the efforts. The US indictment says one unnamed contributor received $400,000 a month for making four videos a week. The FBI also seized 32 web domains, many designed to look like genuine media outlets including Der Spiegel, Le Monde, Fox News and The Washington Post.
Meanwhile, research firm Graphika says a Chinese state-sponsored campaign has been stepping up its impersonation of US voters to spread divisive narratives about sensitive social issues. The research shows the traffic failed to gain much traction, with the exception of a fake media outlet on TikTok that has received 1.5 million views to date.
Although US officials have pointed to China and Russia as the key threats to election integrity, Iran has also been busy - though in its case its efforts appear designed to undermine not only US democracy but also Donald Trump’s re-election campaign.
And one other election note. According to Fox Business, Amazon’s Alexa assistant has been endorsing Kamala Harris and refusing to provide reasons to vote for former President Trump. “This was an error that was quickly fixed,” an Amazon spokesperson told The Register*.*
Threats
Sextortion: A variation on the vile scam now involves including a photo of the target’s home in an effort to make threats to publish compromising videos more convincing. The photos appear to be taken from online services including Google Street View. Krebs on Security
Deepfakes: Palo Alto Networks warns that scam campaigns are using deepfake videos featuring the likeness of various public figures, including CEOs, news anchors, and top government officials. Their aim is to promote too-good-to-be-true investment schemes.
HTML: Crooks are targeting Windows users with information stealing software that’s distributed in emails containing fake purchase orders and invoices. They come in the form of an HTML email with a download link. Fortiguard
Zoom: A convincing scam uses a fake Zoom portal to lure victims into downloading ScreenConnect software. Once installed, ScreenConnect enables an attacker to control the victim’s device without their knowledge. Cyble
AI
Crime: More evidence of the extent to which generative AI is being exploited to create text for fraudulent websites. Netrcraft says it’s seen a surge in such use this year, with a 175% increase in July alone. Fake online shops are a particularly convincing use of AI
Pricing: Design platform Canva has pointed to new AI features as the reason for radical increases in the price of some of its subscriptions. TechCrunch reports that some US users are seeing pricing plans for Canva Teams rise by nearly 320%.
Credentials: Researchers at OpenAI, Microsoft, MIT, and Harvard have come up with a proposed method to verify that someone online is human, without revealing their actual identity. “Personhood credentials” would require physical verification at trusted locations and create an anonymised proof to confirm authenticity online. MIT
Fined: Facial recognition firm Clearview AI is facing a €30.5 million fine in the Netherlands for creating what the country’s data protection regulator says is an “illegal database” of billion of photos of faces. AP
Police: Despite their enthusiastic use of facial recognition, it seems some police officers are less ready for it to be directed at them. Las Vegas police are refusing to provide photos for facial recognition at NFL games in the city, citing the risk of the data falling into the wrong hands and being used against them. Las Vegas Review Journal
In brief
Telegram: Following the arrest of its CEO Pavel Durov, the company has changed its policy and is to start moderating private chats. Durov said an increase in the number of users had made it easier for criminals to abuse the platform. Hmm.
Yubikeys: The bad news; it turns out that there’s a vulnerability in hardware security keys that are considered the most secure authentication method. The good news; it’s really hard to exploit which means only nation states are likely to be able to use it. For anyone at risk of such attacks, it’s worth looking at Bleeping Computer’s report.
Styling: Details of almost five million users of UK online shopping service Lookiero have appeared online, The data comes from a breach in March and includes names, email addresses, phone numbers, and physical addresses. Haveibeenpwned
TfL: Transport for London is continuing to deal with a cyber attack that has forced it to suspend some services and shut down access to digital payment accounts. Details remain sketchy but there are reports that the incident involves a ransomware attack. Computer Weekly
Space: Chinese plans to launch more than 15,000 low earth orbit satellites lay “the groundwork for expanding its digital governance model far beyond its borders,” according to the Australian Strategic Policy Institute. The Register
Spyware: A former member of the Polish government has been charged with diverting legitimate funds to purchase Pegasus spyware. The charges against Michał Woś are part of a wide-ranging investigation into the use of spyware by the former Law and Justice (PiS) government.
Tesla: Police in California’s Bay Area have begun scanning crime scenes to see if any Tesla vehicles are parked nearby. The San Francisco Chronicle says they’re even seeking warrants to seize them to make sure they have access to any video evidence recorded by the cars.
Brat: Charli XCX is an unlikely standard bearer for the importance of secure passwords but then she was forced to scrap an entire album in 2016 after a hacker broke into her computer and leaked unreleased material. She even referenced it in a song called ‘Guess,’ a remix of which was released last month. So, as Eskenzi PR points out, perhaps the least expected aspect of this summer’s “brat” cultural phenomenon is its embrace of cybersecurity. (In case you’re unfamiliar with “brat” then don’t worry. Charli XCX has just said “brat summer” is over anyway.)
Updates
Firefox: Version 130 brings a bunch of new features including improvements to translation, a picture-in-picture function, and an AI chatbot browsing assistant.
Proton: A new VPN protocol (‘Stealth’) is designed to bypass internet censorship by making traffic look like HTTPS, ie web browsing.
Zyxel: Updates for a wide range of its business routers to address a critical vulnerability.
VMware: Update for Fusion Middleware.
Cisco: Fix for security flaw in Cisco’s Identity Services Engine (ISE), for which an exploit is publicly available.
SecureDrop: Client 0.13.2 is a bug fix release that addresses an issue where a failed download would prevent additional downloads from completing until the client is restarted.