Every prompt teaches AI, including the ones with your personal data.
Artificial intelligence tools are becoming part of everyday life. From writing emails and designing logos to editing photos and analysing data, AI is fast, powerful, and incredibly convenient. But as AI becomes more embedded in how we work and live, there’s one question more Australians need to stop and ask: What am I actually giving this tool? Because while AI can feel harmless, the reality is that many people are unknowingly uploading sensitive data in AI systems every day.
AI feels private, but isn’t
AI tools are designed to feel conversational and safe, which makes it incredibly easy to overshare without a second thought. But behind that ease of use, many platforms store prompts and uploaded files, and may use that data to train and improve future models – meaning that what you share doesn’t simply disappear when you close the chat. Once your information is uploaded, you often lose visibility and control over where it’s stored, how long it’s retained, who can access it, and whether it’s reused in ways you didn’t intend. In some cases, that data may also be processed or stored outside Australia, adding another layer of complexity around privacy laws, data protection, and your ability to have that information removed.
What is sensitive data?
When people think of “data,” they often picture spreadsheets, databases, or long, boring documents. But in reality, sensitive data can include everyday information that people upload without a second thought, especially when they’re moving fast or trying to get a quick answer.
This can be as simple as typing out your full name or pasting internal business documents to summarise. Individually, these actions may feel harmless. Collectively, they can expose far more about you than you realise.
Examples of sensitive data you should avoid putting into AI:
- Personal information such as your full name, address, phone number or date of birth.
- Financial information like tax and bank details.
- Login details or internal business documents, including systems, processes or client records.
- Medical or health-related information.
- Images of your face, which is biometric data.
- Voice recordings, including audio notes or samples.
- Identity documents such as licences and passports.
The hidden risk of deepfakes and AI-generated faces
One of the more worrying developments in AI is the rapid improvement of deepfakes and AI-generated imagery, which are becoming increasingly realistic and harder to detect. When you upload a photo of your face, you’re not just sharing a single image, you may be contributing to datasets that allow AI systems to learn, map, and replicate your unique features. This means your likeness could be used to generate entirely new images or videos that look convincingly like you, even if you never created or approved them. In some cases, these images can be altered, manipulated, or placed into misleading or harmful contexts, blurring the line between what’s real and what’s fabricated.
What makes this particularly concerning is that biometric data, like your face, is not something you can change like a password. Once it’s out there, it can be reused, remixed, and redistributed in ways that are difficult to track or control. As AI tools become more accessible, the potential for misuse grows, from impersonation and scams to reputational damage. This shifts the conversation beyond basic privacy concerns into long-term identity protection, where a single uploaded image can have ongoing and unpredictable consequences.
This trend uses ChatGPT to turn photos of real people into playful, stylised caricatures, showing off their hobbies and personality.
At the same time, social media trends are actively encouraging people to hand over this data without much thought. From viral prompts that turn your selfie into an AI avatar or caricature, or generate alternate versions of yourself, these trends gamify the process of uploading your face and personal data. Experts have already warned that these seemingly harmless trends can expose sensitive biometric information and increase the risk of identity theft or deepfake misuse. In many cases, people participate for entertainment or curiosity, without fully understanding where their image is stored, how it may be used, or who might ultimately have access to it.
Continue to use AI, just smarter.
Convenience shouldn’t come at the cost of your privacy, or your business. A little caution goes a long way in keeping your data safe while still leveraging AI to get things done. Before you hit “enter,” pause and ask yourself: does this prompt expose sensitive information?
How you can stay safe while still enjoying the benefits of AI:
- Remove personal names – Instead of typing “My client Sarah wants X,” try “My client [Name] wants X.”
- Blur faces or sensitive images – If you’re sharing photos for editing or analysis, make sure faces or identifiable features are obscured.
- Swap real details for placeholders – Replace addresses, emails, or financial info with generic placeholders like [City], [Email], or [Amount].
AI can be an incredibly powerful tool, and there’s no need to avoid it altogether. The key is using it with intention — understanding what you share and where it goes. By staying mindful of sensitive, personal, and biometric data, you can take advantage of AI’s benefits while protecting your privacy and security.