AI enhances everyday workflows by helping Institutions analyze evidence, surface insights, and work more efficiently across the platform. This article answers common questions about how HelioCampus ensures that AI is used responsibly and effectively. To learn more about HelioCampus’s approach to AI, the HelioCampus Approach to AI blog post is available for review. AI functionality is available in the platform:
How does HelioCampus ensure data privacy and security?
HelioCampus uses the OpenAI API for generative AI functionality. OpenAI API ensures that data is not stored and will not be used to train their AI model. For additional information, these OpenAI resources can be reviewed.
OpenAI Services Agreement
How is HelioCampus generating the information behind the scenes? For example, do you use a Large Language model?
HelioCampus uses the OpenAI LLMs, GPT-4o, and GPT-4o-mini..
What is the process for updating the AI models, and how frequently are they updated?
HelioCampus continually improves the implementation and utilization of the LLM model to enhance specific functionality and implement improvements. OpenAI controls the update process and scheduling for these models rather than HelioCampus. HelioCampus uses the models but does not determine their update frequency or development cycle.
How does the platform ensure the AI behaves ethically and avoids bias?
OpenAI manages all aspects of ethics and bias mitigation. HelioCampus utilizes the models, while OpenAI develops and maintains the fundamental bias prevention mechanisms as part of its model development process.
Are there any system requirements to ensure optimal performance?
No. AI uses the exact system requirements as the platform.
How often is AI data updated?
AI data processing will follow the DBT (Data Build Tool) schedule, which typically runs twice daily for most clients and is relative to the client’s time zone.
Can AI detect stale or broken URLs?
AI has the potential to identify when URLs have gone stale, allowing them to be updated; however, this feature isn't currently implemented.