When Your AI Assistant Becomes a Security Risk

You probably start your morning the same way many Malaysian business owners do: by opening ChatGPT to draft a reply to a difficult customer, asking Claude to summarise a supplier contract, or using Perplexity to research a new market. These tools have quietly become part of your daily operations. But here’s the uncomfortable question — what would happen if someone else was reading those conversations?

AI platform accounts are just as vulnerable to hacking as any other online service. Hackers can break into your ChatGPT, Claude, or Perplexity account, read your business plans, steal customer information, and even impersonate you. A recent TechCrunch guide walks through exactly how to check if your AI accounts have been compromised — and it’s worth your attention before you trust your next prompt with sensitive company data.

What Happened

TechCrunch’s cybersecurity desk published a practical guide on how to tell if your AI platform accounts have been hacked. The guide covers the three most popular AI platforms: ChatGPT, Claude, and Perplexity. For each one, it explains how to spot suspicious logins and what to do if you find one. The advice follows the same pattern as securing any other online account, but with a twist: AI platforms hold something more valuable than your password — they hold your conversation history.

For ChatGPT, the guide points you to your account settings. Click your username in the bottom-left corner, go to “Settings,” then “Security and Login,” and finally “Active Sessions.” There, you’ll see every device currently logged into your account. If something looks unfamiliar, you can log out that single device or click “Log out all” to be safe. If you need to change your password, you’ll have to log out entirely, then use the “Forgot password” option to receive a six-digit code by email and set a new password, as detailed in the TechCrunch walkthrough.

Claude works differently because it doesn’t use passwords at all — instead, it sends a login link to your email. You can still check your “Active sessions” under Settings, then Account. Hover over any unfamiliar session, click the three vertical dots, and choose “Log out” or “Terminate.” There’s also an option to log out of all devices. Perplexity, meanwhile, doesn’t show you where you’re logged in. If you suspect a problem, the guide recommends signing out of all sessions and logging back in with a six-digit code sent to your email.

Why This Matters for Malaysian SMEs

Think about what’s actually inside your AI accounts. If you’re like most SME owners in Malaysia, you’ve pasted customer names, phone numbers, and complaints into ChatGPT to draft responses. You’ve asked Claude to compare quotes from different suppliers. You’ve used Perplexity to research competitors. All of that is sitting in your AI account’s conversation history — and if a hacker gets in, they don’t need to break into your email or your bank. They just need to read your chat logs. The TechCrunch guide notes that these platforms are targeted just like any other online service, and the damage from a breach can be far more subtle than a stolen credit card number.

The risk goes beyond privacy. A compromised AI account can be used to impersonate you. Imagine a hacker logging into your ChatGPT and generating a message to a client that looks exactly like your writing style, asking them to “verify their payment details” or “confirm their bank account number.” Your customers trust you, and they’ll trust a message that appears to come from your business. This is a real attack vector that Malaysian SMEs need to take seriously — not because you’re a big target, but because you’re an easy one. Many small businesses in Malaysia still share passwords across team members or reuse the same password for multiple services, which makes a single breach spread quickly.

The guide also reminds us of basic hygiene: use unique passwords stored in a password manager, and turn on multi-factor authentication (MFA) wherever it’s available. According to the TechCrunch article, ChatGPT and Perplexity both offer MFA, while Claude doesn’t because it relies on email login links instead. For a small business, password reuse across platforms is one of the most common ways accounts get compromised — a breach on one site gives hackers the keys to all your other accounts.

“Your AI accounts are a window into your business operations. If you wouldn’t leave your office door unlocked, don’t leave your ChatGPT session unprotected.”

The Bigger Picture

As Malaysian businesses accelerate their adoption of AI tools, the security industry is watching a new trend: AI accounts are becoming prime targets for attackers. The same reasons you love these tools — they’re fast, they have context, they remember everything — are the reasons hackers love them too. An AI account contains not just your data, but a model of how you think and write. That’s a goldmine for social engineering attacks, because a hacker can study your tone, your vocabulary, and your typical requests to craft messages that are nearly impossible to distinguish from your own.

The good news is that the platforms are responding. ChatGPT and Perplexity now offer MFA, and all three platforms give you some way to review or terminate active sessions, as shown in the TechCrunch guide. But security is a shared responsibility. The real lesson for Malaysian SME owners is to build a simple routine: check your active sessions regularly, enable MFA on every account that supports it, and never reuse passwords. Make it a team policy, not an afterthought.

Here’s a quick reference you can share with your staff:

Platform	How to check for suspicious logins	What to do if compromised
ChatGPT	Settings → Security and Login → Active Sessions	Log out the device, or log out all and reset your password
Claude	Settings → Account → Active Sessions	Terminate the session or log out of all devices
Perplexity	No session list available	Sign out of all sessions, then log in again with email code

Take ten minutes this week to check your AI accounts. It’s a small investment of time that could save you from a very awkward conversation with a client whose data was exposed. In the age of AI, your conversations are your business — protect them like it.

Ready to Streamline Your Operations?

Technology moves fast. Your operations should keep up. AutoRunBiz builds AI systems that run your daily workflows — from WhatsApp order capture to accounting. Book a free 15-min ops audit →