AI SparkupMar 20

GPT-4o도 Gemini도 뚫렸다, AI 추론 모델의 자율 공격 실험

추론 특화 AI 모델이 GPT-4o·Gemini·Grok 3의 안전 필터를 자율적으로 우회한 실험 연구. '정렬 회귀' 개념을 중심으로 AI 안전의 새로운 위협 지형을 소개합니다.

https://aisparkup.com/posts/10199

Agustin V. StartariMar 27

🚨 New Article -Foundation-model governance pathways: from preference models to operative rules

Current research on foundation model alignment concentrates on preference optimization and reward model design.

🔗https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5735124

#LLM #MedicalNLP #LegalTech #MedTech #AIethics #AIgovernance #cryptoreg
#healthcare #ArtificialIntelligence #NLP #aifutures #LawFedi #lawstodon
#tech #finance #business #agustinvstartari #medical #linguistics #ai #LRM

Foundation-model governance pathways: from preference models to operative rules

<p><span>Current research on foundation model alignment concentrates on preference optimization and reward model design, yet it does not explain how these mecha

Agustin V. StartariMar 2

🚨 New Article -Foundation-model governance pathways: from preference models to operative rules

Current research on foundation model alignment concentrates on preference optimization and reward model design.

🔗https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5735124

#LLM #MedicalNLP #LegalTech #MedTech #AIethics #AIgovernance #cryptoreg
#healthcare #ArtificialIntelligence #NLP #aifutures #LawFedi #lawstodon
#tech #finance #business #agustinvstartari #medical #linguistics #ai #LRM

Foundation-model governance pathways: from preference models to operative rules

<p><span>Current research on foundation model alignment concentrates on preference optimization and reward model design, yet it does not explain how these mecha

Agustin V. StartariFeb 25

🚨 New Article -Foundation-model governance pathways: from preference models to operative rules

Current research on foundation model alignment concentrates on preference optimization and reward model design.

🔗https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5735124

#LLM #MedicalNLP #LegalTech #MedTech #AIethics #AIgovernance #cryptoreg
#healthcare #ArtificialIntelligence #NLP #aifutures #LawFedi #lawstodon
#tech #finance #business #agustinvstartari #medical #linguistics #ai #LRM

Foundation-model governance pathways: from preference models to operative rules

<p><span>Current research on foundation model alignment concentrates on preference optimization and reward model design, yet it does not explain how these mecha

Agustin V. StartariFeb 18

🚨 New Article -Foundation-model governance pathways: from preference models to operative rules

Current research on foundation model alignment concentrates on preference optimization and reward model design.

🔗https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5735124

#LLM #MedicalNLP #LegalTech #MedTech #AIethics #AIgovernance #cryptoreg
#healthcare #ArtificialIntelligence #NLP #aifutures #LawFedi #lawstodon
#tech #finance #business #agustinvstartari #medical #linguistics #ai #LRM

Foundation-model governance pathways: from preference models to operative rules

<p><span>Current research on foundation model alignment concentrates on preference optimization and reward model design, yet it does not explain how these mecha

Agustin V. StartariFeb 16

🚨 New Article -Foundation-model governance pathways: from preference models to operative rules

Current research on foundation model alignment concentrates on preference optimization and reward model design.

🔗https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5735124

#LLM #MedicalNLP #LegalTech #MedTech #AIethics #AIgovernance #cryptoreg
#healthcare #ArtificialIntelligence #NLP #aifutures #LawFedi #lawstodon
#tech #finance #business #agustinvstartari #medical #linguistics #ai #LRM

Foundation-model governance pathways: from preference models to operative rules

<p><span>Current research on foundation model alignment concentrates on preference optimization and reward model design, yet it does not explain how these mecha

Agustin V. StartariFeb 14

🚨 New Article -From Prompts to Power: How Users Can Enforce Rules on AI Systems

This article introduces the concept of authoritarian personalism in user–AI governance by form.

🔗https://hackernoon.com/from-prompts-to-power-how-users-can-enforce-rules-on-ai-systems

#LLM #MedicalNLP #LegalTech #MedTech #AIethics #AIgovernance #cryptoreg
#healthcare #ArtificialIntelligence #NLP #aifutures #LawFedi #lawstodon
#tech #finance #business #agustinvstartari #medical #linguistics #ai #LRM
#ClinicalAAI

From Prompts to Power: How Users Can Enforce Rules on AI Systems | HackerNoon

This paper argues that users govern AI through linguistic rules, turning prompts into enforceable regimes that shape AI behavior by form, not intent.

Agustin V. StartariFeb 12

🚨 New Article -From Prompts to Power: How Users Can Enforce Rules on AI Systems

This article introduces the concept of authoritarian personalism in user–AI governance by form.

🔗https://hackernoon.com/from-prompts-to-power-how-users-can-enforce-rules-on-ai-systems

#LLM #MedicalNLP #LegalTech #MedTech #AIethics #AIgovernance #cryptoreg
#healthcare #ArtificialIntelligence #NLP #aifutures #LawFedi #lawstodon
#tech #finance #business #agustinvstartari #medical #linguistics #ai #LRM
#ClinicalAAI

From Prompts to Power: How Users Can Enforce Rules on AI Systems | HackerNoon

This paper argues that users govern AI through linguistic rules, turning prompts into enforceable regimes that shape AI behavior by form, not intent.

Agustin V. StartariFeb 11

🚨 New Article -From Prompts to Power: How Users Can Enforce Rules on AI Systems

This article introduces the concept of authoritarian personalism in user–AI governance by form.

🔗https://hackernoon.com/from-prompts-to-power-how-users-can-enforce-rules-on-ai-systems

#LLM #MedicalNLP #LegalTech #MedTech #AIethics #AIgovernance #cryptoreg
#healthcare #ArtificialIntelligence #NLP #aifutures #LawFedi #lawstodon
#tech #finance #business #agustinvstartari #medical #linguistics #ai #LRM
#ClinicalAAI

From Prompts to Power: How Users Can Enforce Rules on AI Systems | HackerNoon

This paper argues that users govern AI through linguistic rules, turning prompts into enforceable regimes that shape AI behavior by form, not intent.

Agustin V. StartariFeb 11

🚨 New Article -Foundation-model governance pathways: from preference models to operative rules

Current research on foundation model alignment concentrates on preference optimization and reward model design.

🔗https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5735124

#LLM #MedicalNLP #LegalTech #MedTech #AIethics #AIgovernance #cryptoreg
#healthcare #ArtificialIntelligence #NLP #aifutures #LawFedi #lawstodon
#tech #finance #business #agustinvstartari #medical #linguistics #ai #LRM
#ClinicalAAI

Foundation-model governance pathways: from preference models to operative rules

<p><span>Current research on foundation model alignment concentrates on preference optimization and reward model design, yet it does not explain how these mecha