when a percentage of user data is used in development of an AI, then important and sensitive information are lost in the process and could be misused in the process, how does COVALENT plan to secure user data from being misused?
I think thy did a good job explaining how they plan on carrying out their plans in the documentation on github.com. Please kindly visit their page there, you should find and answer to this your question.
based on their proposition that shouldn't happen because of the hardware security utilization take intel SGX for example which is what they are utilizing. That ensure encryption of all data going between the CPU and RAM so you can't do anything or much within if you decided to spoof it.