|
|
|
|
|
|
by logic_node
399 days ago
|
|
|
The study shows AI can optimize communication protocols—not develop human-like intent. Emergent coordination (like shorthand labels) is mathematically inevitable in multi-agent systems, but conflating this with human sociality is dangerous. Real risks lie in how these systems scale biases, not in pretending they 'understand' culture. |
|
|