4 ms·
> Generative AI models like OpenAI’s ChatGPT and Google’s Gemini can now generate realistic text and images that are often indistinguishable from human-authored
by cebert 2y ago
> Generative AI models like OpenAI’s ChatGPT and Google’s Gemini can now generate realistic text and images that are often indistinguishable from human-authored content
We are fortunate that it currently is possible to identify AI generated content. It’s generally written too well, repeats clauses and phrases, and doesn’t always sufficiently summarize.
I’ve observed a growing trend of using AI to generate comments on LinkedIn posts. This technique is presumably utilized to increase impressions or followers. Currently, it’s very obvious when the comments are AI generated.
- Bluestein 2y agoI think the issue here is that - what you well place as tells that are "obvious" to us when identifying AI content ... ... are more "instinctual" tells than anything else, and are hard to hard-commoditize in automated or algorithmic ways that can be applied at the same scale AI can generate content, and do so with certainly - or, sufficient, faithful certainty that the content is AI-generated.-