More substantial changes include adding two new properties to robots.txt specifically for AI training, which would allow content owners to express more granular control over how AI systems use their content. It would give publishers and content owners more control over how AI systems use their content, including the ability to limit snippet length, require attribution, and allow or deny use for AI training. While there’s momentum and enthusiasm behind creating technical solutions, many of the thorniest issues—like how to interpret the lack of signal, combine conflicting signals, or manage different preferences—are far from resolved.The working group agreed to support both opt-in and opt-out signaling for AI usage, allowing content owners to explicitly declare their preferences.