AI點解成日「扮嘢」?OpenAI話有原因!

Ai




OpenAI指大型語言模型(LLM)因獎勵機制而誘發「幻覺」現象

人工智能公司OpenAI最近發表一份研究報告,指出大型語言模型(Large Language Models,LLM)訓練的聊天機械人因為獎勵機制,傾向於猜測答案,而非坦白未知,這就是所謂的「幻覺」現象。

報告中提到,這些「幻覺」源自一種二元分類錯誤,即模型在將新資訊歸類為兩個選項之一時出錯。LLM被優化成為優秀的「考試答題者」,在不確定時猜答案反而能提升整體表現,類似學生在選擇題考試中猜答案,因為留白得分為零,猜即使錯也有機會得分。

LLM的運作機制是按得分計算:正確答案得分,空白或說「不知道」則不獲得分數。這種設計導致模型傾向於提供答案,即使不確定也會嘗試猜測。

這份報告發表之際,OpenAI剛推出了新一代GPT-5,聲稱該模型比前代GPT-4o的錯誤率降低46%,更接近「無幻覺」狀態。但根據美國NewsGuard的最新研究,ChatGPT系列模型仍有40%的回答包含錯誤資訊。

AI面對「無法回答」問題時的困境

通過預先訓練和後期調整,聊天機械人學習在大量文本中預測下一個詞語。OpenAI指出,語言模型在拼寫和文法等有明確規範的任務上表現良好,但面對某些題材或資料類型,準確分類仍有困難甚至不可能。

例如,模型能夠準確分辨標籤為「貓」或「狗」的圖片,但若按寵物生日作分類,則無法準確識別。這種任務無論算法多先進,錯誤率始終存在。

報告強調,模型永遠無法達到百分百準確,因為現實世界中某些問題本質上就是「無法回答」的。

為減少幻覺,OpenAI建議用戶可指示模型在不確定時回答「我不知道」,並調整其得分機制,鼓勵模型誠實表達未知,而非盲目猜測。

編輯評論與深入分析

OpenAI此次研究揭示了現有大型語言模型核心運作的「隱性問題」:為了達到更高的表面準確率,模型被設計成在不確定時也要「硬答」,這種機制從根本上誘發了「幻覺」問題。對用戶來說,這種現象顯得像是聊天機械人「說謊」或「編故事」,但其實是系統獎勵機制反映出的結果。

這提醒我們,AI並非真實「懂得」答案,而是在有限範圍內計算概率和得分。若想提升AI的可靠性,除了技術層面改進外,更需要重新設計激勵機制,讓模型學會在不確定時坦白「不知道」,這不僅有助於減少錯誤資訊的傳播,也能建立用戶對AI的信任。

此外,報告提到某些問題本質上無法由AI解答,這是非常重要的認知界限。未來AI應與人類專家合作,對這類問題給予明確提示,避免誤導。

最後,OpenAI推出的GPT-5雖有提升,但根據獨立研究仍有相當比例的錯誤,顯示技術進步固然重要,監管和用戶教育同樣不可忽視。AI的發展不能只追求「表面準確度」,更需重視「透明度」和「誠實度」,這是打造負責任和可信賴AI的關鍵。

以上文章由特價GPT API KEY所翻譯及撰寫。而圖片則由FLUX根據內容自動生成。

🎨 Nano Banana Pro 圖像生成器|打幾句說話就出圖

想畫人像、產品圖、插畫?SSFuture 圖像生成器支援 Flux Gemini Nano Banana Pro 改圖 / 合成, 打廣東話都得,仲可以沿用上一張圖繼續微調。

🆓 Flux 模型即玩,不用登入
🤖 登入後解鎖 Gemini 改圖
📷 支援上載參考圖再生成
⚡ 每天免費額度任你玩
✨ 即刻玩 AI 畫圖
Create a hyper-realistic 8K close-up body portrait of a female model, using the uploaded photo as the exact facial reference. Maintain 100% accuracy of the facial features — do not alter or modify any aspect of the face. Render the skin texture, lighting, and overall composition with photo-realistic detail, ensuring lifelike color tones and natural depth of field. A man with his original hair is sitting casually on a white cube, smiling warmly at the camera. He is wearing a cream-colored cable-knit sweater, blue jeans, and brown loafers. His legs are crossed, with one hand resting on his knee.
The background reveals a cozy and festive living room. A large, beautifully decorated Christmas tree with numerous warm lights and gold ornaments stands prominently behind him. Several wrapped gift boxes are visible at the base of the tree. To his left, another smaller decorated Christmas tree and a wreath on the wall further enhance the holiday atmosphere. The lighting is soft and inviting, creating a warm and welcoming scene. 上身近鏡 Create a hyper-realistic 8K close-up body portrait of a female model, using the uploaded photo as the exact facial reference. Maintain 100% accuracy of the facial features — do not alter or modify any aspect of the face. Render the skin texture, lighting, and overall composition with photo-realistic detail, ensuring lifelike color tones and natural depth of field. A man with his original hair is sitting casually on a white cube, smiling warmly at the camera. He is wearing a cream-colored cable-knit sweater, blue jeans, and brown loafers. His legs are crossed, with one hand resting on his knee.
The background reveals a cozy and festive living room. A large, beautifully decorated Christmas tree with numerous warm lights and gold ornaments stands prominently behind him. Several wrapped gift boxes are visible at the base of the tree. To his left, another smaller decorated Christmas tree and a wreath on the wall further enhance the holiday atmosphere. The lighting is soft and inviting, creating a warm and welcoming scene.