Enterprise Secure AI vs. Open AI
Yes, there are significant risks associated with using Open AI in the everyday work environment, especially when it’s used by employees who are untrained or unaware of AI security practices.
However, it’s important to understand that when you place an AI platform in a silo (an enterprise account) for security reasons, it will have limited data for training and development, and it will soon stop evolving (becoming “smarter”)—especially with restricted usage. “Limited” here refers to the variety of usage, prompts, areas of application, and the number of users and inputs. It will no longer be refined and developed at the same speed as Open Generative AI, which uses the entire world as its training field. As a result, the quality of output may become poor and underwhelming.
I recommend carefully evaluating both the risks and opportunities of using confined enterprise AI platforms and Open AI. This evaluation should not be made for the company as a whole but for each specific department or team. For example, the sales and marketing team, which works with publicly known information and materials, may benefit more from training on how to use Open AI securely, rather than limiting AI’s language capabilities to a confined enterprise model.
