高質量英語演講&專訪影片哪裡看?
請您點選上方“精彩英語演講”,並“設為星標”
全網最新的英語演講&專訪第一時間為您奉上


隨著大模型已成為 AI 產業發展的主導力量,一場學術界和 OpenAI 之間的思辨已經開始。
作為連續舉辦五屆的重要學術性會議,智源大會已經有超過 500 點陣圖靈獎得主的頂尖專家參與。今年則包括圖靈獎得主 Geoffrey Hinton," 圖靈獎 " 得主、" 深度學習三巨頭 " 之一的楊立昆(Yann LeCun),OpenAI CEO 薩姆 · 奧爾特曼(Sam Altman),中國科學院張鈸院士等 AI 領域重磅人士,共議人工智慧前沿與熱點話題。
6 月 10 日,OpenAI 創始人 Sam Altman 以影片連線的方式現身於中國北京舉辦的 2023 智源人工智慧大會,這是 Altman 首次對中國觀眾發表演講,會上其大談 AI 風險與未來通用 AI 發展模式,直言為了弄清楚通用 AI 技術,OpenAI 必須要推進 AI 研究變革。

Altman 在演講中引用了《道德經》,談及大國合作,表示 AI 安全始於足下,各國間必須進行合作和協調。
Altman 認為,人工智慧爆發式增長的背景下,未來 10 年就可能會出現超強 AI,呼籲全球共同監管,並且在相關的研究部署上對齊,建立全球信任,而 Altman 之所以持續呼籲協同監管,一方面是因為 AI 全球大爆發,另一方面則是因為 " 最小化風險 " 的理念。
對於大模型在實際應用場景中可能存在的錯誤,Altman 的看法是 " 希望在風險較低的時候犯錯 ",然後透過迭代來讓其變得更好,以此來闡釋 OpenAI 的人工智慧迭代理念。
提及超強 AI,此前未來生命研究所曾發起呼籲 " 暫停 GPT-4 以上大模型研發 6 個月 " 的公開信,OpenAI 在這塊的動作,包括後續 GPT-5 的規劃一直是業內關注的焦點,今天的線上問答環節,在被張宏江問及了 AGI 的未來,以及是否會很快看到 GPT-5 時,Altman 稱自己也不確定,但明確表示 GPT-5 的誕生," 不會很快 "。
關於 OpenAI 是否會提供開源大模型時,Altman 稱 " 會有很多開源大模型 ",但同樣沒有具體的時間表。

↓↓↓ 上下滑動,檢視演講稿 ↓↓↓

Sam Altman, the CEO of OpenAI, has called for a global cooperation in setting up the regulatory guardrails for powerful artificial intelligence (AI) systems at an industry forum held in Beijing on Saturday, noting that China can play a significant role with some of its "best AI talent in the world."
"China has some of the best AI talent in the world… given the difficulties solving alignment for advanced AI systems requires the best minds from around the world," Altman said at an industry forum hosted by the Beijing Academy of Artificial Intelligence (BAAI) on Saturday.
Altman predicted that in the next decade, artificial general intelligence systems, commonly called AGI, may surpass human expertise "in nearly every domain," and could eventually exceed the collective productivity of some of the largest companies.
"The potential upside here isn't one. AI revolution will create shared wealth and make it possible to dramatically improve the standard of living for everyone. We must manage the risk together in order to get there," Altman said.
At present, a new round of AI boom triggered by ChatGPT is sweeping the world including China. While people have been thrilled for the productivity improvement that AI can bring, they are increasingly aware of the huge risks, with governments slowly stepping in to harness the powerful yet dangerous tool.
In May, hundreds of industry leaders, including Altman himself, have signed a letter warning that "mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war." Altman made a point to reference the IAEA, the United Nations nuclear watchdog, as an example of how the world came together to oversee nuclear power.
"We need international cooperation to build global trust in the safe development of increasingly powerful AI systems. In a verifiable way, I have no illusions that this will be easy. We will need to devote significant and sustained attention as an international community to get this right," Altman said.
In a possible indication to the current geopolitical conflict's impact on the global AI cooperation, Altman noted that "great powers may have their share of differences." This is true today, as it has been before. But even during the most difficult times, great powers have found ways to cooperate on what matters. "We see great potential for researchers in the US, China, and around the world to work together to achieve the shared goal."
Altman used an ancient Chinese proverb, "a journey of a thousand miles begins with a single step," in describing how the world should start on the difficult but necessary cooperation.
In talking about the next frontier that OpenAI is engaging on in a later Q&A session, Altman said OpenAI is discussing about opening source more models in the future, but does not have a specific model or timetable.
"Certainly, at some point, we'll try to do a GPT-5 model in the sequence, but not anytime soon and we don't know exactly when," Sam said, adding that he's hoping someday the company can get to return to robotics – a sector the company has worked on at the very start.
The two-day conference hosted by the BAAI has invited leading figures in the industry such as AI "godfather" Geoffrey Hinton and Turing Award winner Yann André LeCun. Apart from exchanging on AI frontiers and technology progress, this year's conference placed a special focus on AI regulation and alignment.
Industry players and experts at the forum have also emphasized on the importance of global cooperation in harnessing the "double-sword" technology, which is especially urgent as the large models are development "too fast."
"We are fully aware of the urgency of AI regulation, but under the current global situation, as some in the US are hyping China threat with AI as a key battleground, global cooperation and an international rule could be hard, though some cooperation might be achievable in certain areas,"Wang Peng, a research fellow at the Beijing Academy of Social Sciences, told the Global Times on Saturday at the sidelines of the event.
Looking ahead, the development of large models and major technological breakthroughs is clearly be related to the life and death of human society, Wang said, noting that it's critical for some nations to drop closed and retrogressive minds of "small courtyards and high walls," join forces together with the rest of the world, and concentrate on overcoming difficulties for the benefit of all mankind.

防止再次失聯,請立即關注備用號


想第一時間觀看高質量英語演講&採訪影片?把“精彩英語演講”設定為星標就對了!操作辦法就是:進入公眾號——點選右上角的●●●——找到“設為星標”點選即可。



