1.6M Exposed by a Phone Call 📞 and the First AI Attack That Ran Itself 🤖

Two stories today show how fast the ground is shifting: 1.6 million people had their personal details exposed because someone talked their way past a help desk, and researchers documented the first cyberattack that largely ran itself. Here's what actually matters.

📞 One Convincing Phone Call, 1.6 Million People Exposed

RingCentral — the cloud phone and collaboration platform used by thousands of businesses — confirmed a breach after attackers talked their way into its systems through what the company called a "sophisticated social engineering campaign." No zero-day exploit, no malware: just a human being convinced to hand over access. The stolen data, now catalogued by Have I Been Pwned, includes names, email addresses, physical addresses, and phone numbers for roughly 1.6 million people.

Here's the "so what": that combination of details is exactly what scammers need to make their next approach sound legitimate. Expect a wave of convincing phishing emails and phone calls referencing real addresses and real accounts. It's also a reminder that your security is only as strong as the person answering the phone — which is why identity verification procedures for help desks and IT support have quietly become one of the highest-leverage controls a business can implement.

Read more →

🤖 The First Cyberattack That Mostly Ran Itself

Researchers at Dream have identified what they believe is the first publicly known "near-autonomous" AI cyberattack — a campaign against Taiwanese government targets that stole more than 2,500 personnel records. The attackers, suspected to be linked to China, built their malware on open-source AI agent frameworks and let it run self-directed "learning cycles," where the software analyzed its own failures, adapted, and pushed deeper — eventually reaching IT supply chains, nuclear safety agencies, and energy companies.

The important caveat: researchers noted the campaign still required significant human engineering to work. This isn't a robot hacker acting alone. But it is a preview of where things are heading — attacks that iterate at machine speed and don't get tired, bored, or discouraged. The defensive takeaway is uncomfortable but simple: if attackers can probe continuously, then detection and response can't be something you check on Monday mornings.

Read more →

Stay ahead of threats with GOCO Security at gocosecurity.com.