There are obvious data privacy risks here. Sometimes, I forget that not everyone works in the legal industry and is hyper-aware of confidentiality and data privacy, the way you are when that is your firm’s business, but this is not one of those times. Every business, including Microsoft, would be unhappy if a user synced company data into their consumer OneDrive account.
The pressure to become AI experts and make the organization’s rollout of AI tools secure only adds more mental load to already overloaded cybersecurity professionals.
I probably haven’t been thinking enough about how the budget for AI might come at the cost of several other IT projects and what that might mean in terms of old, insecure, and inefficient technology not being replaced.
As I have said before, if your data is stored somewhere outside your control, it’s only a matter of time before it gets hacked. Your AI assistant will have a lot of private information, making it a prime target.
This makes me wonder if eDiscovery productions shouldn’t be done using a solution like SharePoint, with permissions that allow the opposing side to read documents online but not download a copy. Thus, they would not have a copy to upload to a GenAI platform.
I also understand why this happens, though. Even out-of-the-box AI tools like Copilot or ChatGPT require some level of data access and security. That’s a complicated task that, in normal times, would require a months-long project, at minimum.
Custom AI projects that require sharing data across your environment and allowing the model access to all that data in a way that shuts everyone up about what they can’t do with AI will lead to mistakes. We’re seeing them, just like when people rush other IT projects under fire to implement technology they don’t have time to understand.