"Emily M. Bender, Timnit Gebru, Angelina McMillan-Major, and Margaret Mitchell introduced the stochastic parrot framing as a criticism of large language models."
The landmark FAccT paper introduced the stochastic parrot framing while examining the assumptions and consequences of scaling language models. The authors questioned whether benchmark improvements and increasing model size should be confused with genuine language understanding and called for closer examination of the risks and research priorities involved.
View Source