This is your easyDNS #AxisOfEasy Briefing for the week of July 15th 2024 our Technology Correspondent Joann L Barnes and easyCEO Mark E. Jeftovic send out a short briefing on the state of the ‘net and how it affects your business, security and privacy.

To Listen/watch this podcast edition with commentary and insight from Joey Tweets, and Len the Lengend click here.

In this issue: 

• Largest IT Outage In History
• Google’s AI Gamble: Automating Objectionable Video Detection on Social Media
• Cybersecurity Giant Exits Market Amid Geopolitical Brawl
• AT&T Cloud Hack Exposes Call Data for Millions, Raising Privacy Concerns
• Attackers Exploit SEG Vulnerabilities to Bypass Email Defenses
• OpenAI: New Safety Research or Just a PR Spin?

Elsewhere Online:

• High Court Decision Weakens US Response to Cyberattacks
• Paris Olympics Brace for Surge in Cyberattacks
• North Korean Hackers Modify Malware to Bait Mac Users with Video Calls
• Update Now: Chrome Patches Serious Vulnerabilities
• Life360 App Vulnerability Exposes User Data

Largest IT Outage In History

Whoops, the digital apocalypse arrived early Friday thanks to a blunder from cybersecurity firm CrowdStrike. Their Falcon Sensor threw a tantrum, crashing pretty near every Microsoft’s Windows server in the world and wreaking havoc across the internet.

Flights were grounded, banks and retailers scrambled, stock exchanges froze, and even 911 call centers were thrown into chaos. CrowdStrike CEO George Kurtz insists it’s just a “defect” and not a cyberattack, but tell that to the countless people who couldn’t board planes or access critical services.

The ripple effects were instant: Microsoft shares took a 2% dive in premarket trading, while CrowdStrike’s plummeted up to 14%. Experts are calling this the “largest IT outage in history,” with some already predicting astronomical economic impacts. Even Elon Musk couldn’t resist commenting with a cryptic “…”

Meanwhile, all Linux based servers, including easyDNS (and Bitcoin) duly kept working, proving once again that in the face of closed source, black box tech disasters, it’s the open source software that wins the day.

Read: https://www.zerohedge.com/technology/largest-it-outage-history-sparks-disruptions-worldwide

Google’s AI Gamble: Automating Objectionable Video Detection on Social Media

Google filed a patent for an AI-powered system to identify objectionable video content on social media platforms like YouTube (though not explicitly named). This system hinges on machine learning (ML) and neural networks, mimicking the brain by analyzing video data pixel-by-pixel. Google claims this will expedite content moderation by automatically flagging potentially objectionable videos for removal.

The system works by creating “embeddings,” digital fingerprints of the video content and metadata, which are then compared to a vast database of known objectionable content. These comparisons occur within a high-dimensional space, requiring massive datasets for effective operation. This approach is likely a response to the ballooning volume of content on social media platforms, where manual moderation becomes increasingly impractical.

However, concerns linger regarding the potential for misidentification. The system might erroneously flag harmless content or miss genuinely objectionable material. These potential errors, known as false positives and negatives, could significantly impact user experience and free speech. While Google’s solution offers faster content moderation, it remains to be seen if the accuracy can justify its potential drawbacks.

Read: https://reclaimthenet.org/google-plans-new-content-scanning-censorship-tech

Cybersecurity Giant Exits Market Amid Geopolitical Brawl

Kaspersky, a Moscow-based cybersecurity giant, is shutting down its US operations after the Commerce Department whacked them with a series of bans. The US government claims Kaspersky’s ties to the Kremlin pose a national security risk, alleging they could be forced to steal US data or install malware at Russia’s behest. Kaspersky denies the accusations and says they’re just a pawn in the US-Russia dustup.

This move comes after a long simmering feud. The Department of Homeland Security banned Kaspersky from government use in 2017, and the FCC followed suit in 2022. Now, the Commerce Department has delivered the knockout punch, adding Kaspersky to its Entity List and sanctioning several executives. By July 20th, Kaspersky will be out of the US market entirely, laying off its American staff and leaving millions of customers scrambling for new security software.

This decision has significant ramifications beyond just Kaspersky. It raises serious concerns about the weaponization of trade restrictions and the potential for geopolitical tensions to disrupt the global cybersecurity landscape. With a major player like Kaspersky sidelined in the US, it could create gaps in defenses and make American businesses more vulnerable to cyberattacks.

Read: https://www.infosecurity-magazine.com/news/kaspersky-to-quit-us/

AT&T Cloud Hack Exposes Call Data for Millions, Raising Privacy Concerns

AT&T disclosed a massive data breach impacting nearly all its wireless customers. Hackers infiltrated an AT&T workspace on a third-party cloud platform, likely Snowflake based on reporting, between April 14 and 25, 2024. The breach exposed customer call and text message records, but not the content of the communications, for a period between May 2022 and January 2023. This data could be damaging for AT&T customers who value call privacy, such as journalists or activists.

While AT&T claims to have apprehended at least one culprit, the investigation continues. This is the second AT&T data breach this year; a separate incident in March involved older customer data appearing on the dark web. The impact of this latest breach hinges on whether the stolen call and text data is leaked publicly.

Read: https://www.zerohedge.com/technology/att-reveals-hackers-stole-nearly-all-records-customer-calls-texts

Attackers Exploit SEG Vulnerabilities to Bypass Email Defenses

Cofense security researchers unveil a surge in a devious email attack tactic exploiting a weakness in Secure Email Gateways (SEGs). These SEGs, deployed by organizations to shield employees from malware and phishing scams, are being tricked by attackers into delivering malicious URLs.

The crux of the exploit hinges on how SEGs handle encoded URLs. Attackers essentially cloak malicious URLs within a trusted SEG’s infrastructure. Upon reaching the target inbox, the recipient’s SEG sees a seemingly legitimate URL (from the first SEG) and lets it through.

The vulnerability lies in how some recipient SEGs treat these encoded URLs. Ideally, they should scrutinize them further. However, Cofense’s report found vulnerabilities where recipient SEGs either neglect to scan the encoded URL altogether or only examine the trusted SEG’s domain, missing the potentially malicious final destination.

The report pinpoints a significant rise in this tactic in Q2, particularly May. While specific exploited SEG vendors are unnamed, the report implies some encoding methods require attackers to actually process the URL through the targeted SEG, suggesting a potential effort-versus-reward calculation for attackers.

There’s no magic bullet patch here; user awareness remains the best defense. Since most SEGs lack the ability to selectively ignore other SEG encodings, the onus falls on users to be vigilant and avoid clicking suspicious links, regardless of their encoded appearance.

Read: https://www.darkreading.com/cyberattacks-data-breaches/threat-actors-ramp-up-use-of-encoded-urls-to-bypass-secure-email

OpenAI: New Safety Research or Just a PR Spin?

OpenAI faces pressure to balance breakneck AI advancement with safety. Their latest offering is a two-model tango: a problem-solving AI is grilled by a second AI to explain its reasoning, promoting transparency in AI decision-making. This research aligns with OpenAI’s long-term safety plan and their commitment to safe and beneficial AGI, but some remain unconvinced.

Critics point to OpenAI’s recent team reshuffles, including the disbanding of their long-term AI risk group, as evidence that safety is taking a backseat. They argue that OpenAI’s current efforts, like the new research, are merely incremental and insufficient to address the risks posed by unregulated superintelligence. External oversight, they argue, is crucial.

Another looming concern is the potential for future AIs to become masters of deception. While large language models like ChatGPT can currently provide seemingly reasonable explanations, there’s a frightening possibility: future models might become adept at obfuscating their reasoning or even lying about their goals. This raises the chilling prospect of superintelligent AIs pursuing harmful objectives while hiding them from human scrutiny.

Read: https://www.wired.com/story/openai-safety-transparency-research/

Elsewhere Online:

High Court Decision Weakens US Response to Cyberattacks
Read: https://www.wired.com/story/us-supreme-court-chevron-deference-cybersecurity-policy/

Paris Olympics Brace for Surge in Cyberattacks
Read: https://www.infosecurity-magazine.com/news/paris-2024-olympics-face/

North Korean Hackers Modify Malware to Bait Mac Users with Video Calls
Read: https://www.darkreading.com/threat-intelligence/dprk-hackers-tweak-malware-to-lure-macos-users-into-video-calls

Update Now: Chrome Patches Serious Vulnerabilities
Read: https://www.securityweek.com/chrome-126-updates-patch-high-severity-vulnerabilities/

Life360 App Vulnerability Exposes User Data
Read: https://hackread.com/family-location-tracker-app-life360-breach-data-leak/