§ Controversies
Claims in active dispute. Each controversy has multiple positions, each with proponents and sources. ‘Uncontested’ is not a requirement — scholarship evolves.
Position A
“AGI is likely within the next 3-5 years (by ~2028-2030), and current scaling and reasoning approaches are sufficient to…”
Position B
“AGI is 10-20+ years away and requires fundamental breakthroughs beyond current paradigms”
Position A
“Training AI models on publicly available content is transformative fair use that benefits society”
Position B
“Using copyrighted works to train AI without permission or compensation is infringement at scale”
Position A
“Safety Through Engagement — AI labs should collaborate with militaries to shape responsible use from within”
Position B
“Normalization of Militarization — Lab engagement legitimizes military AI and erodes ethical red lines”
Position A
“Open weights democratize AI, enable safety research, and prevent dangerous concentration of power”
Position B
“Above a capability threshold, open release creates unacceptable proliferation risks that cannot be mitigated…”
Position A
“Scaling continues to produce meaningful gains — apparent plateaus are temporary and will be overcome by next-generation…”
Position B
“Training-time scaling is hitting diminishing returns, but test-time compute (reasoning) and efficiency improvements…”