Instead of crafting solutions for hypothetical harms, we advocate focusing on already existing and very significant safety issues–namely AI’s reliance on personal data & the vulnerabilities created by foundation models' use in nat’l security. 2/

→ View original post on X — @mer__edith