Are there any LLMs that were trained solely on content/data gathered with the creators’ consent?
I’m a former website content writer. I found out that my former employer used my past work (including some writing from as far back as the 2010s) to build a custom GPT to replace me. They then generated a bunch of SEO content with hallucinated “information”, which I find all the more horrifying. It’s a really shitty feeling to know you and your work have been used like that.
I was already uneasy with the most popular LLMs, knowing that they were trained on many authors’ and artists’ work who would have preferred otherwise, but this really does it. I don’t want to use any tools that have been trained on someone’s work against their wishes.
So. Are there any LLMs that are solely trained on writing/work that was gathered with the creators’ explicit consent? If not, is someone building one? (And if not, please, for the love of God, someone build one!)