FFT news digest — 5 May 2023
FFT news digest — 5 May 2023
The week in AI
Cyber criminals are as entranced by ChatGPT as everyone else, according to Meta, which says it has blocked more than 1,000 unique malicious web addresses with an AI theme. The scams often involve mobile apps or browser extensions posing as ChatGPT tools. Some of them do provide a measure of ChatGPT functionality but all of them aim to steal account credentials. Meta says people who manage businesses on Facebook or use the platform for work have been particular targets because their accounts often have a linked credit card. Meanwhile, ChatGPT was forced offline following a cyber attack apparently caused by a vulnerability in the its programme code. The issue revealed the chat history of active users and made payment information visible . The issue was quickly fixed but it does underline the need to apply the same - or more - caution to ChatGPT as to any other online platform or service (as evidenced by the experience of a Financial Times journalist who was spied on by TikTok.) In other AI news; - More alarm sirens from key figures in AI, including the now-departed head of AI at Google …and the head of AI at the Pentagon who said he was “scared to death” about how people might use ChatGPT and other consumer-facing AI agents. - A Wall Street Journal columnist used various AI tools to create a clone that could trick her bank’s biometric system, and her family. - NewsGuard found 49 websites created by generative AI and posing as news outlets with generic names like News Live 79. - A new AI system can read people’s minds by translating their brain activity. University of Texas
Threats
Phishing: The latest phishing survey suggests it continues to be a growing blight for individuals and organisations. Zscaler says (R) the number of phishing attacks rose 47% in 2022 compared to the previous year. The US and the UK were the most targeted countries, with education, finance and government seeing the highest number of attacks. Zscaler says Microsoft, OneDrive and Finance were the top three most impersonated brands - though other surveys frequently give the top place to DHL. Firefighter : A widespread campaign is trying to fool people into donating money to firefighting associations. The criminals are abusing PayPal to send what look like genuine invoices and billing reminders. The campaign is focused on the US at the moment but the same tactic has been used around the world. Avanan Password Managers : A new version of information stealing software is targeting a broader range of apps, including 1Password and Keepass password managers. It hides inside fake software activators and the like. Trend Micro Cookies : Facebook says it shut down a campaign in which hackers were stealing the cookies used to authenticate a user’s browsing session. The criminals used information stealing software that masqueraded as browser extensions, adverts and ChatGPT assistants. Checkout : Some of the most ingenious frauds involve hijacking the online shopping experience. The latest way of doing this uses realistic-looking payment forms that sit on top of the genuine ones. As Malwarebytes explains , the fake pages are so realistic that there’s no way to distinguish them from the real thing. The only protection is to use a credit card for online payments so that you don’t end up footing the bill for any fraudulent charges.
Passwords
It’s so boring to bang on about passwords that we try to avoid it but there are more signs this week sees that their days are numbered. In a blog post, Google says users can now use passkeys to access personal accounts. This means an end to passwords and two-step verification. Instead, a passkey allows the user to sign in to their account by unlocking their computer or mobile device with a fingerprint, face recognition or PIN/passphrase/pattern. Apple already supports this approach on Ventura and iOS 16, and a number of online services are adopting it. The approach is (for the moment) far, far more secure than any other authentication mechanism. It’s well worth using.
Nation states
New reports provide some useful insights into the latest tactics used by nation state hackers. In a detailed look at Iran’s activities, Microsoft says Teheran has accelerated its cyber-enabled influence operations in attempts to bolster Palestinian resistance and sow fear in Israel, foment unrest in Bahrain, and counter the ongoing normalisation of Arab-Israeli ties. New Iranian influence techniques include the use of SMS messaging and victim impersonation. Meanwhile, Ukraine says Russia appears to have been using fake Windows updates to attack government organisations. The scale of China’s cyber resources was the subject of a US congressional hearing. The head of the FBI said Chinese hackers outnumber FBI cyber personnel by at least 50 to 1.
In brief
Spyware : Amnesty International says Pegasus spyware was found on the personal device of one of the Dominican Republic’s most prominent investigative journalists. Researchers say her smartphone was infected three times between 2020 and 2021. HTTPS : That padlock in Chrome’s address bar is about to disappear - mainly, it seems, because so many people misunderstand what it means. As we’ve explained countless times in our training, it doesn’t mean a website is safe or genuine, it just denotes that the connection is secure. Google will replace it with a different symbol that most likely will also be misunderstood. Tracking : Two men accused of stealing $1.1 million in an armed robbery in Chicago were tracked down thanks to an Apple AirTag. Indeed Bluetooth trackers are so effective that New York city is handing out 500 AIrTags in a bid to reduce car crime. At the same time, malicious use of the trackers has led Apple and Google to cooperate on creating an industry standard to address unwanted tracking.