Understand Your Company's AI Policy
Before you start using any AI tool for work, your first step should be to check your company's official policy. Many organisations have already established clear guidelines on which AI platforms are approved and what types of data are considered too sensitive
to use with them. These policies are not meant to restrict innovation but to create a secure framework for it. They often specify whether you should use a business-tier account (like ChatGPT Team or Microsoft Copilot), which typically comes with contractual promises that your data won't be used for public model training. If your company doesn't have a policy, ask for one. Raising the question helps management understand the need for clear rules.
Never Input Sensitive or Confidential Information
This is the golden rule of using public AI. Treat any information you enter into a free or public AI tool as if you were posting it on a public forum. Sensitive data includes a wide range of information: personally identifiable information (PII) of customers or employees, financial records, unannounced product details, proprietary source code, legal documents, and internal strategy memos. Pasting this kind of information into a public AI can lead to it being stored on third-party servers and potentially used to train the model, making it part of a massive dataset accessible by others.
Anonymise and Generalise Your Prompts
You can still get valuable assistance from AI without revealing specifics. The key is to practice good 'prompt hygiene'. Instead of pasting a full, sensitive document and asking for a summary, anonymise the content first. Replace real names, company details, and specific figures with generic placeholders like "[Client Name]," "[Product X]," or "[Q4 Revenue Figure]." For example, rather than asking the AI to 'critique the marketing plan for our new product launching in Delhi,' ask it to 'critique a marketing plan for a new tech product launching in a major metropolitan city.' This gives the AI enough context to provide a useful framework without exposing your company’s secrets.
Use Approved, Enterprise-Grade AI Tools
Companies are increasingly adopting enterprise-level AI solutions that are designed with security and privacy in mind. These tools are often hosted on private infrastructure or come with robust data protection agreements that ensure your company’s data remains your own. Microsoft Copilot for Microsoft 365, for instance, is built to respect your organisation's existing security and privacy settings. If your company provides access to such a tool, always use it instead of a public, free version for work-related tasks. These platforms are evaluated by IT and legal teams to ensure they meet compliance standards like GDPR and other data protection laws.
Critically Review All AI-Generated Content
AI tools can 'hallucinate,' meaning they can invent facts, statistics, or sources that sound plausible but are entirely incorrect. Never assume that AI-generated output is factually accurate or ready for use without a thorough review. Before incorporating any AI-generated text, code, or data into a report, presentation, or client communication, fact-check it against trusted sources. You remain responsible for the accuracy and quality of your work, and blindly trusting an AI's output can damage your credibility and your company's reputation.
Be Aware of How Your Data Is Used
Many free AI services explicitly state in their privacy policies that they may use user inputs to train and improve their models. Some platforms offer an option to opt out of having your data used for training, but you may have to actively look for this setting. Before using a new tool, take a moment to understand its data retention and usage policies. This small step can prevent a major data leak down the line. Using secure networks, like a VPN, and keeping your devices updated also adds another layer of protection when interacting with any online service, including AI.
