The software program business has began creating an enormous array of synthetic intelligence (AI) functions based mostly on giant language fashions (LLMs). Whereas many safety threats to LLMs are just like these affecting conventional software program, LLMs and their functions additionally face distinctive safety dangers because of their particular traits. These dangers can typically be mitigated or diminished by making use of particular safety structure patterns. Listed here are 10 methods to mitigate and scale back safety dangers in LLM functions.1. Establish, authenticate and authorize all of the principalsThis contains people and brokers that p