Atoosa Kasirzadeh Calls AI Safety Language Encrypted

Atoosa Kasirzadeh, Staff Research Scientist at Google DeepMind and CMU professor on leave, posted that part of the AI safety community has developed an almost encrypted language. She said discussing its claims with scientists and people outside the space often requires citing blog posts and mastering an insular vocabulary. Kasirzadeh added that she sometimes gets lost in the communication and called for insisting on clarity.