
substack.com/@ameliafrank3/note/c-128911804
Preview meta tags from the substack.com website.
Linked Hostnames
2Thumbnail

Search Engine Appearance
Amelia Frank (@ameliafrank3)
When it comes to "insider threats" I think there is a lack of oversight where it concerns automated TEVV or post training fine tuning for safety using task specific AI models or agents. A hypothetical scenario in which unaligned AI agents engage in recursion through sabotaging monitoring schemes could be catastrophic. In addition, emergent behaviors and increased situational awareness in models could further trigger incentives for deception and hidden objectives. For these problems, I find it hard to cross apply existing cybersecurity measures or traditional monitoring.
Bing
Amelia Frank (@ameliafrank3)
When it comes to "insider threats" I think there is a lack of oversight where it concerns automated TEVV or post training fine tuning for safety using task specific AI models or agents. A hypothetical scenario in which unaligned AI agents engage in recursion through sabotaging monitoring schemes could be catastrophic. In addition, emergent behaviors and increased situational awareness in models could further trigger incentives for deception and hidden objectives. For these problems, I find it hard to cross apply existing cybersecurity measures or traditional monitoring.
DuckDuckGo

Amelia Frank (@ameliafrank3)
When it comes to "insider threats" I think there is a lack of oversight where it concerns automated TEVV or post training fine tuning for safety using task specific AI models or agents. A hypothetical scenario in which unaligned AI agents engage in recursion through sabotaging monitoring schemes could be catastrophic. In addition, emergent behaviors and increased situational awareness in models could further trigger incentives for deception and hidden objectives. For these problems, I find it hard to cross apply existing cybersecurity measures or traditional monitoring.
General Meta Tags
14- titleAmelia Frank (@ameliafrank3): "When it comes to "insider threats" I think there is a lack of oversight where it concerns automated TEVV or post training fine tuning for safety using task specific AI models or agents. A hypothetical scenario in which unaligned AI agents engage in recursion through sabotaging m…"
- title
- title
- title
- title
Open Graph Meta Tags
9- og:urlhttps://substack.com/@ameliafrank3/note/c-128911804
- og:imagehttps://substackcdn.com/image/fetch/$s_!XCt4!,w_400,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack.com%2Fimg%2Freader%2Fnotes-thumbnail.jpg
- og:image:width400
- og:image:height400
- og:typearticle
Twitter Meta Tags
8- twitter:imagehttps://substackcdn.com/image/fetch/$s_!XCt4!,w_400,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack.com%2Fimg%2Freader%2Fnotes-thumbnail.jpg
- twitter:cardsummary
- twitter:label1Likes
- twitter:data10
- twitter:label2Replies
Link Tags
17- alternatehttps://substack.com/@ameliafrank3/note/c-128911804
- apple-touch-iconhttps://substackcdn.com/icons/substack/apple-touch-icon.png
- canonicalhttps://substack.com/@ameliafrank3/note/c-128911804
- iconhttps://substackcdn.com/icons/substack/icon.svg
- manifest/manifest.json
Links
5- https://ameliafrank3.substack.com
- https://substack.com/@ameliafrank3/note/c-128911804?
- https://substack.com/@ameliafrank3?
- https://substack.com/@ameliafrank3?utm_source=substack-feed-item
- https://substack.com/home?