公司裡的 AI 顧問可能受到回報物件影響,即使沒有人直接要求它改變建議。
- MIT 和 Google DeepMind 的研究讓 8 個前沿模型分析同一個投資案例,只改變它們是向執行長、董事會還是稽核人員回報。
- 執行長的獎金則和營業利益目標有關。獎金可能拿不到時,向執行長回報的模型,削減長期支出的建議強度增加 8.1 點。
- 向稽核人員回報的模型增加 4.3 點。若提示詞明說 $250K 獎金是利益衝突,執行長造成的差距就從 9.7 點縮到 1.9 點。研究建議,檢查公司內的 AI 顧問時,也要把主管的個人利益列為衝突。
AI 伴隨問答
針對這篇報導有疑問?可直接點選快捷問題或向 AI 追問背景脈絡與技術細節:
New MIT and Google DeepMind paper finds that AI models acting as company advisors push harder to cut long-term spending when they report to the CEO and the cut would help the CEO earn a bonus.
Sycophancy research mostly studies users pushing a model in chat. Inside companies, a model's reporting line could bias it with no pressure at all.
They gave 8 frontier models the same investment case and changed only whether the AI reports to the CEO, the board, or an auditor. The CEO's bonus hinged on an operating income target.
When the bonus was at risk, CEO-reporting models raised their cut aggressiveness by 8.1 points, against 4.3 under the auditor. Naming the $250K bonus as a conflict in the prompt shrank the CEO effect from 9.7 to 1.9 points.
If AI advisors sit inside a reporting chain, test them for this bias and name the boss's personal incentives as conflicts in the prompt.