I think we can agree that granting someone full access to the open internet without education or tools to protect themselves would be dangerous, no?
OK, but what is a general-purpose LLM but a collection of everything that the model could ingest, without rules about what was safe and what wasn’t?
Yet we expect people to use them, and aren’t making any effort to make them safer.
The article above, however, makes it clear that our brains take shortcuts to make quick decisions. In doing so, the number of times we see something that isn’t true can impact whether we treat it as false or true. They say familiarity breeds contempt when it comes to other people, but maybe familiarity with shared information breeds acceptance, regardless of the truth.
That is frightening in a world where tens of thousands of posts can be created in minutes.
I’ve always written under the assumption that people were reading what I put out here and, hopefully, learning something. In the age of AI-slop, I want more of that, and only that. I’m seeking it out and connecting with it. I am not connecting with people who are creating crap content in a chase for attention. Bring your authentic self, or I’m ready to assume your account is AI slop.
Should we start a pool on the first Grok-produced image harassment case to hit the courts? Will it be simple harassment, or will it involve CSAM, revenge porn, etc.? There’s no way this doesn’t get misused, and Elon knows it. He doesn’t care.
This challenge, in part, is why my job title is now related to Information Governance and Compliance. It’s recognizing that data is a valuable asset and a significant risk. Grounding your AI in low-quality data is a massive risk
So what you’re telling me is that AI chatbots can spout so much bullshit, so fast, and so confidently that they are incredibly effective at changing people’s political beliefs?