The next time someone targets your CEO, they most likely won’t be a nation-state operative or a professional cyberattacker. They may be a disgruntled customer with a ChatGPT subscription and 15 minutes of free time.

This is no exaggeration. This is the new norm.

Large language models (LLMs) such as ChatGPT have dramatically changed the equation for who can find executive PII and how quickly. The protective barrier that previously made your personal data difficult to access is essentially gone. Anyone with a grudge and an internet connection can now surface details that once required specialized knowledge.

The Uncomfortable Reality

Here’s what security leaders need to understand: AI tools pull their information from publicly available sources on the open web. Data broker sites, people-search databases, social media profiles, public records. These tools don’t have a secret back door to private information. They’re just extraordinarily good at aggregating what’s already out there.

That distinction matters. Because it means the fix isn’t blocking AI. The fix is to reduce the pool of information that AI can draw from in the first place.

The most dangerous combination security teams should worry about? Home addresses linked to valid phone numbers. That pairing creates an actionable path for anyone with bad intentions. It’s specific enough to be useful and verifiable enough to be trusted. When AI can surface that combination in seconds, the time between someone deciding to cause harm and having the information to do it shrinks to nearly nothing.

This Isn’t About Nation-State Actors Anymore

For years, we talked about sophisticated adversaries when discussing executive threats. State-sponsored hackers. Professional cybercriminals. People with resources and training.

Those threats haven’t gone away. But AI has democratized the reconnaissance phase of an attack. A disgruntled former employee, an angry investor, a random person who saw your executive on the news and decided they didn’t like what they heard. Any of these people can now do in minutes what used to take a trained investigator days or weeks.

A recent study found that nearly half (45%) of corporate end users have used a generative AI platform in their workflows. Nearly 40% of uploaded files contain personally identifiable information or payment card data. Organizations have been leaking sensitive information to these systems, sometimes unknowingly.

The problem is compounded by the fact that AI does not discriminate between correct and incorrect answers. Generative AI systems provide answers based solely on your query; if it cannot find sufficiently accurate information, it will still produce an answer. Therefore, threat actors may rely on AI-generated false information; however, they can also produce true information with inaccuracies that are difficult to verify.

What Actually Works

The best option for removing your personal information from the internet is to contact the data brokers directly and have it removed at the source. Opting out of individual AI-based tools provides some relief; however, it does not eliminate the risk of being identified by other applications that use the same or similar data found in search engine results and data-broker databases.

The second component of maintaining control over how your personal information is used is continuous monitoring of where executive information appears online. Personal information rarely stays deleted indefinitely. Data brokers continually populate their databases. While a single cleaning effort may provide a false sense of security, you will need continuous oversight of where your information appears online, and a quick response mechanism in place so that, if/when it is again available, you can act quickly to protect yourself.

Additionally, as technology is increasingly used to identify and locate individuals, protecting the spouses and children of corporate executives has become essential. Spouses and children are overlooked threat vectors. For example, school directories, sports teams, and social media connections can serve as conduits for collecting and aggregating information about an individual that can be queried with alarming efficiency.

The Speed Problem

The part that security teams tend to underestimate: the gap between threat identification and threat action is collapsing.

In the past, if someone wanted to find your CEO’s home address, they’d spend time searching, verifying, and cross-referencing. That friction created time. Time for them to reconsider. Time for their anger to cool. Time for security teams to potentially identify concerning behavior before it escalates.

AI compresses that timeline. The information is available in seconds. The verification happens instantly. By the time someone has decided to act, they already have what they need.

This means security posture needs to shift from reactive to preemptive. Waiting until a threat materializes to assess your executive’s digital exposure is no longer viable. By then, a threat actor already has the information. The only effective strategy is to reduce that exposure before anyone comes looking.

The Questions That Matter Now

AI-enabled reconnaissance is happening today. Executives are being researched, profiled, and potentially targeted using tools that make the process trivially easy. The question isn’t whether your organization should respond. It’s whether you’re responding fast enough.

The fundamentals haven’t changed: control the information, control the risk. What’s changed is the urgency. Every day that executive PII remains accessible on data-broker sites is another day that information can be weaponized by anyone who seeks it.

Security leaders should be asking hard questions. Do we know where our executives’ personal information is exposed? Are we monitoring for new exposure? How quickly can we respond when sensitive data surfaces?

The organizations that answer those questions now will be better positioned when AI-enabled threats become more sophisticated. Those who wait will find themselves playing catch-up against adversaries with all the time in the world to research and none of the friction that used to slow them down.