|
Post by juthi52943 on Jan 4, 2024 5:32:55 GMT
For clouddelivered LLM models , terms of use and privacy policy again become a key issue as they are for a publicly released ChatGPT bot, but are more likely to be aligned with existing cloud terms of service. Organizations need to understand how to manage the data that will be used to finetune or extend prompts. Will the data be available to researchers or vendor partners. If so, in what form? Will they be shared separatelyor in aggregation Job Function Email List with them? Under what conditions will a suppliers employee be able to view inquiries? Private LLMs are likely to be very expensive. However, after a security assessment, they may be suitable for handling organizational data. How ChatGPT makes life easier for cybercriminals. Several experiments have already been conducted and reported showing how ChatGPT can help write malware. The concern is that ChatGPT could be used by someone with bad intentions but not enough skills to create tools that they would not be able to implement without the model. The current pain point is that ChatGPT produces answers that seem convincing regardless of whether they are correct.
|
|