As Generative AI is adopted by businesses that are keen to offer new or competitive service, it is important that security features are not ignored.
Access control refers to letting the right users in a business have access to different levels of the organization. For example, financial information can be read by the finance team and the owners of a business, and sales information should be kept secure from competitors.
Letting AI gain access to information runs the risk of information leakage. A user or a client who is not given to certain information when using the normal information systems of an organization, may be able to bypass such ‘access control’ restrictions by simply asking an AI agent or chatbot to provide such information.
At the same time, just removing all such information from being read by your AI chatbot or agent is not a viable solution, as it restricts what the AI may do for many legitimate scenarios.
Workmaster is built to support fine-grained access control to both users – and the AI agents and chatbots that work on their behalf. This means the AI has access to the same information that the user herself would have were she to perform that action directly in the Workmaster-built business application.
Workmaster supports role-based access, as well as ‘row-based’ access for data where users may only access their own information.
Imagine an AI driven ‘self-help’ app used to record medical information. Users may access their own information, but not those of others – this ownership-based data access is enforced by Workmaster so that users and their bots only access their own records. Its impossible for a Workmaster chatbot to answer questions related to another patient’s medical records – because the chatbot simply can’t read such records.
Such fine-grained access control will be a mainstay of future AI led IT systems. You can just start using them in Workmaster, today.