Yes, but one of the conflicting interests is illegal. We all know these companies pirate a huge amount of copyrighted data to train their LLMs and VLMs. Clear copyright infringement, Anthropic just lost a few billion dollars for this.

In addition, the training process attempts to reproduce the copyrighted training data as perfectly as possible, with the intent to rent the resulting model out for commercial gain afterwards. Many argue that this is not fair use, but another instance of copyright infringement.

And if the previous infractions weren't enough, OpenAI's customers are now generating mass videos of copyrighted characters.

So, while it may be common corporate speak, it is still snake-tongued weasel-blather that downplays the illegality of their actions.