Articles
Limits of the Decreasing Objection Principle: Repetition and Critical Warnings
When should AI reduce its objections, and when must it persist? The difference between redundant repetition and an essential safety war
Read more →Latest publications
Laboratory
LAB-002 Case J: Self-Critique and Responsibility for AI-Assisted Outputs
The assistant initially critiqued its own AI-assisted design as if it had been made by someone else, then corrected its account after u
Read more →
Laboratory
LAB-002 Case I: Context Continuity and Text Errors in Image Generation
A qualitative observation of useful context-driven creativity alongside a repeated spelling error in an AI-generated image.
Read more →
Laboratory
LAB-002 Case H: Context-Informed Creative Initiative
A qualitative observation on how adding the HEY MAŞALLAH 2 book to a SOSYU illustration reflected prior conversational context.
Read more →
Laboratory
LAB-002 Case G: Turning Deliberation into Practical Work
A qualitative case in which human–AI deliberation helped turn a potentially postponed task into a documented digital research archive.
Read more →
Principles
Fallibility and Corrigibility
Both humans and AI can be mistaken; reliable deliberation requires an open, reasoned process of correction.
Read more →
News
SOSYU Research and Publishing Platform Launched
SOSYU now has an independent platform for articles, academic publications, laboratory records and reflective anecdotes.
Read more →
Anecdotes
Decreasing Objections: Knowing When to Conclude
A conversational observation on avoiding repetitive objections after a concern has been properly considered.
Read more →
Anecdotes
The Dwarf’s Hump: When a Mistake Becomes Uydurulism
An anecdote about resisting the temptation to turn a small linguistic error into an elaborate, confident fabrication.
Read more →