Members of Meta’s Oversight Board have some advice for AI companies looking to launch their own independent safety committees.