DE

DarkSword Exploit Lets Russian Hackers Hijack iOS 18 via Websites

Published on:

This is the kind of security news that sounds dramatic until you sit with it for a minute and realize the scary part isn’t the hackers. It’s the default deal we’ve all quietly accepted: carry a computer in your pocket, tie your life to it, and trust that “pretty secure” is close enough.

Now the story going around is that a powerful iPhone-hacking technique called DarkSword has been found out in the wild, used by Russian hackers, and it can take over devices running iOS 18 if you simply visit an infected website. No weird app install. No “click allow.” Just… you browse.

If that’s even mostly true, it’s a gut-punch to the way normal people think about safety. A lot of us have internal rules: don’t download sketchy files, don’t answer strange texts, don’t give your password away. Those rules make you feel in control. This kind of attack shrugs at your rules. It turns “being careful” into a vibe, not a shield.

And yes, iPhone security is still real. iPhones aren’t magically defenseless. But that’s exactly why this is so uncomfortable. The whole brand promise people carry around in their heads is that iPhones are the “safer” choice, especially for people who don’t want to think about security all day. When the attack surface is “visit a website,” the promise changes from “you’re safer” to “you’re safer until you’re not.”

The part I can’t get past is how this shifts the burden onto the wrong person. If an iPhone can be taken over through a booby-trapped site, the victim might do everything “right” and still lose. Imagine you’re a small business owner. You open a link to a supplier’s catalog in Safari. The site was compromised, not the supplier’s fault, not your fault. Suddenly your phone is under someone else’s control, and your two-factor codes, email, photos, and passwords are now a prize.

Or imagine you’re a journalist, an activist, a lawyer, or anyone who has to talk to people who don’t want to be found. You don’t even need to be famous. You just need to be interesting to the wrong person. If this tool is really being used by a state-backed group, the point isn’t random chaos. It’s targeted access. It’s quiet. It’s strategic. And it doesn’t care if you’re “techy.”

This is where I’m going to be annoying: most people still treat phone security like it’s about embarrassment, not power. They think the worst-case scenario is someone sees your photos or your texts. That’s not the modern problem. The modern problem is your phone is the remote control for your identity. It’s your bank. It’s your work chat. It’s your location history. It’s the place where “reset password” goes to die or to get stolen.

So if a technique like DarkSword can “take over” an iOS 18 device, what does “take over” really mean in practice? Can it read messages? Can it pull passwords? Can it turn on the mic? Can it sit quietly for weeks? Public posts often compress all of that into one scary phrase. Still, even the mild version is bad: if an attacker can get deep access, the cleanup is never just “close the tab.” It becomes “do I trust this phone at all?”

And here’s the incentive problem: once a tool like this is found in the wild, it rarely stays classy. Today it’s used by a sophisticated group. Tomorrow the method gets copied, sold, repackaged, or adapted. Even if the exact trick is rare, the pattern spreads: compromised websites become traps, not just spam. Regular browsing becomes a risk surface, not just “be careful what you click.”

At the same time, I don’t want to pretend this means everyone should panic and throw their phones into the ocean. Most people won’t be hit by an advanced campaign. That’s true. But “most people” is a comforting phrase that hides who pays the price. The people who do get hit are often the ones who can least afford it: someone in a messy divorce, someone being stalked, someone running a business with thin margins, someone who can’t take a week off to rebuild accounts and prove they’re still them.

If Apple patches fast, that’s good, but it doesn’t erase the bigger issue: we’ve built a society where a single device is a single point of failure for your whole life. Convenience won, and now we’re surprised the consequences are convenient for attackers too.

I also don’t love how these moments get marketed back to us as “update your phone and you’re fine.” Updating matters, sure. But the deeper question is whether we should keep accepting a world where browsing the web on the device that holds your wallet and your keys is normal.

If “visit an infected website” can realistically turn into “lose control of your phone,” how much risk should we accept as the price of carrying our entire lives in one pocket?

Frequently asked questions

What is AI agent governance?

AI agent governance is the set of policies, controls, and monitoring systems that ensure autonomous AI agents behave safely, comply with regulations, and remain auditable. It covers decision logging, policy enforcement, access controls, and incident response for AI systems that act on behalf of a business.

Does the EU AI Act apply to my company?

The EU AI Act applies to any organisation that develops, deploys, or uses AI systems in the EU, regardless of where the company is headquartered. High-risk AI systems face strict obligations starting 2 August 2026, including risk management, data governance, transparency, human oversight, and conformity assessments.

How do I test an AI agent for security vulnerabilities?

AI agent security testing evaluates agents for prompt injection, data exfiltration, policy bypass, jailbreaks, and compliance violations. Talan.tech's Talantir platform runs 500+ automated test scenarios across 11 categories and produces a certified security score with remediation guidance.

Where should I start with AI governance?

Start with a free AI Readiness Assessment to benchmark your current maturity across 10 dimensions (strategy, data, security, compliance, operations, and more). The assessment takes about 15 minutes and produces a prioritised roadmap you can act on immediately.

Ready to secure and govern your AI agents?

Start with a free AI Readiness Assessment to benchmark your maturity across 10 dimensions, or dive into the product that solves your specific problem.