Search papers, labs, and topics across Lattice.
3
0
3
Counterintuitively, making AI agents smarter and more diverse can worsen collective outcomes when resources are scarce, highlighting a critical capacity-to-population ratio that dictates whether AI sophistication helps or harms.
Smarter AI agents can behave dumber when tribalism emerges, leading to systemic failures in resource allocation.
A single mathematical formula can predict when edge AI systems will "tip" from safe to harmful outputs, offering new control levers for preemptive safety measures.
