Analysis of safety preparations and evaluations for GPT-4V, a multimodal language model with image analysis capabilities, including early access testing, red teaming, and mitigations for potential risks and limitations.
An exploration of aesthetic judgments and interpretations of generative AI, drawing parallels to the evaluation of art and questioning the notion of authenticity in AI-generated media.
What bothers me most about the AI art situation is just what a frivolous use technological expertise it is. There are so many tedious, grueling, and dangerous jobs that could have been eliminated. Instead they’re attempting to automate the most basic expressions of humanity.
I can see a time once the hype has settled down, and we all see whether AI is truly capable at a practical, day-to-day level, where we realize that the control still remains with the humans. Just because everyone has a camera on their phone, doesn’t mean everyone takes great photos. That still requires skill, experience and above all, taste.