The hidden risks of using chat GPT at work

AI and ChatGPT, they’re the big buzz words and with 700 million weekly active users on ChatGPT, it’s not without reason. More businesses and individuals are turning to ChatGPT to draft emails, come up with ideas and even manage companies’ personal data. It’s a great tool to increase your efficiency when used correctly. 

AI productivity is on the rise but so is the risk. LayerX reported that: 

• • 45% of enterprise employees now use generative AI tools in their daily work. 

• • 77% of those users have copied and pasted company data into chatbot prompts. 

• • 22% of that shared data contains sensitive information including personal and payment data. 

Is this a risk you’re willing to take? 

There are real risks for businesses from both attackers and competitors exploiting generative AI tools. Employees can inadvertently train models with sensitive company data, which may then be exposed externally including on the dark web. For context, over 225,000 OpenAI credentials have already been leaked through malware and other attacks. Here are the key areas you need to monitor: 

• Sensitive data leaks: Employees may unintentionally share confidential business information with ChatGPT, risking exposure and misuse.  

• Transmission vulnerabilities: Data sent to AI platforms can be intercepted by cybercriminals if not properly secured.  

• Account security risks: A compromised ChatGPT account could reveal chat histories containing sensitive or strategic company data.  

• Unintended model training: Information shared may be used to train the AI, potentially benefiting competitors with insights from your business. 

Copilot: The hidden champion... 

The risks of using ChatGPT are real, but the world is moving fast toward AI-driven workflows. So, what options exist that deliver value without exposing sensitive data? One often overlooked solution is Microsoft Copilot an AI assistant that enhances your workflows, streamlines tasks, and strengthens productivity without feeding data back into public models or training competitors’ AI. 

David Keeling Managing Director for Cloud & Security at Intercity sat down with Matt Weston CEO at Vantage to discuss the opportunities Copilot can bring to your business. 

How else can you tighten up your security? 

• Training your people: Educate your employees to understand the risks associated with using ChatGPT and teach how to use AI in a secure manner 

• Deploy security controls for ChatGPT usage: Implement access restrictions, monitoring, and authentication measures to prevent unauthorised use and protect company data. 

• Establish data handing policies: Define what data can and cannot be shared with AI tools, setting clear boundaries to safeguard sensitive information. 

These measures are a step in the right direction. It creates an open conversation channel in your business to discuss and more importantly acknowledge the risks and uses of Open AI. 

Key takeaways... 

Generative AI tools are transforming how businesses work but they’re also creating new entry points for attackers. Sensitive data, intellectual property and personal information are all at risk of exposure through unsecured AI use.  

By putting clear security policies in place for AI tools, your business can harness their benefits, such as Copilot, and confidently unlock innovation without opening the door to data leaks or compliance risks.