At the end of last month, President Donald Trump brought tech giants to Washington for a luncheon and a symbolic signing ceremony for a voluntary safety agreement.
上月底,總統唐納·川普將科技巨頭帶到華盛頓,出席午餐會,並舉行了一場自願性安全協議的象徵性簽署儀式。
He described the agreement as "almost a constitution," centered on "superintelligence"—the government's new preferred term for AI.
他形容該協議「幾乎是一部憲法」,其核心在於「超級智慧」——這是政府對AI的新偏好稱呼。
On Wednesday, in a crowded event space in San Francisco, safety challenges were not the foremost concern on people's minds.
週三,在舊金山一處擁擠的活動空間裡,安全挑戰並非眾人最關心的問題。
At that gathering, founders and funders stated that this hands-off approach by the government would yield higher profits for companies building AI-driven defense and security systems.
在該場合,創辦人與資助者表示,政府這種放手不管的做法,將為打造AI驅動國防與安全系統的公司帶來更高利潤。
At a conference themed "Investing in the Frontier," attendees praised the bullish stance on government support for industry.
在一場以「投資前沿」為主題的會議上,與會者對政府支持產業的看多立場表示讚揚。
The conference was one of a series of pop-up events hosted by the venture capital firm Andreessen Horowitz, whose head serves as Trump's science and technology adviser.
該會議是創投公司Andreessen Horowitz所主辦的一系列快閃活動之一,而該公司負責人為川普的科學與科技顧問。
"I'd like to begin with a figure that I believe truly highlights the opportunities in defence investment, and that is one trillion US dollars.
「我想先從一個我認為真正突顯國防投資機會的數字開始,那就是1兆美元。
For the first time, the US government is allocating one trillion dollars of its defence spending to national defence.
美國政府首次將其國防支出的1兆美元用於國防。
And of that, 13.5 billion dollars will go towards autonomous systems and AI."
而其中135億美元將用於自主系統和AI。」
said Kyle Stanford, a senior venture capital research director at the market intelligence platform PitchBook.
市場情報平台PitchBook的創投資深研究總監凱爾·史丹佛說。
"This clearly brings a substantial amount of capital into defence technology spending, more than we have ever seen before."
「這顯然為國防科技支出帶來了大量資本,比我們以往見過的還多。」
A day later, three researchers at OpenAI's headquarters said that the ChatGPT developer had fired them, and they posted a statement saying that "fear and unclear rules are hindering AI safety work."
一天後,OpenAI總部的三名研究人員表示,這家ChatGPT開發商已將他們解僱,並發文稱「恐懼和不明確的規則阻礙了AI安全工作」。
In addition, David Robinson, a head of OpenAI's safety team who resigned only recently, expressed concern about allowing these advanced laboratories to regulate themselves.
此外,OpenAI安全團隊的一名負責人大衛·羅賓森直到最近才辭職,他對讓這些先進實驗室自我監管表達了擔憂。
"This cannot replace binding legislation, and it effectively grants these companies considerable freedom," Robinson stated.
「這無法取代具有約束力的法律,而且它著實賦予了這些公司相當大的自由。」羅賓森表示。
"The intervals between model generations are measured in weeks... I believe we will face several extremely severe safety challenges in the much nearer future."
「模型世代之間的時間間隔以週為單位……我認為我們將在更近的未來面臨若干極為嚴峻的安全挑戰。」
Trump has staked his political capital on the success of the AI industry, and he has called its unfettered development the key to winning the AI race against China.
川普已將其政治資本押注於AI產業的成功,並稱其不受約束的發展是贏得對中國AI競賽的關鍵。
On the first day of his second term, the Republican president repealed a Biden-era executive order that was partly intended to guard against the risks that AI systems pose to consumers, workers, and national security.
在他第二個任期的第一天,這位共和黨總統廢除了拜登時期的一項行政命令,該命令部分旨在防範AI系統對消費者、勞工和國家安全構成的風險。
His administration has since shifted federal policy away from oversight toward rapid growth, including by rescinding a previous requirement that AI developers share safety test results with the federal government.
他的政府此後將聯邦政策從監督轉向快速成長,包括撤銷先前要求AI開發者與聯邦政府分享安全測試結果的規定。
In December, Trump signed an executive order aimed at blocking state-level AI regulation, saying that a patchwork of rules could stifle the industry.
12月,川普簽署了一項行政命令,旨在阻止州層級AI監管,稱拼湊式的規則可能扼殺該產業。
Concerns about AI safety, focused mainly on existential risks rather than present-day real-world harms, became the focus of debate last month.
對AI安全的擔憂,主要集中於生存性風險,而非當前現實世界的危害,這在上個月成為辯論焦點。
At that time, a relatively unknown Anthropic engineer resigned and called for a pause in technological development to prevent "superhuman" systems from escaping control.
當時,一名相對不知名的Anthropic工程師辭職,並呼籲暫停科技發展,以防止「超人類」系統逃脫控制。
Subsequently, the CEOs of Anthropic and OpenAI also echoed calls to slow down AI development.
隨後,Anthropic和OpenAI的執行長也附和放慢AI發展的呼籲。
To some critics, their public declarations seemed aimed at courting investors ahead of IPOs and the midterm elections; during the midterms, the political winds could shift as many voters oppose data centers.
對一些批評者而言,他們的公開宣言似乎是為了在上市和期中選舉前討好投資人;期中選舉時,隨著許多選民反對資料中心,政治風向可能轉變。
"The only control or 'guardrails' AI needs is a strong and smart (high IQ!) president, and the United States has this in full measure, and then some!" Trump responded on Truth Social, warning voters not to "kill the golden goose" — a term he has used to refer to AI and data centers.
「AI唯一需要的控制或『護欄』,就是一位強大又聰明(高智商!)的總統,而美國完全具備這一點,而且綽綽有餘!」川普在Truth Social上回應道,並警告選民不要「殺掉金雞母」——他曾用這個詞來指AI和資料中心。
This document was jointly signed by representatives of OpenAI, Anthropic, Meta, Google, SpaceXAI, and Nvidia.
本文件乃由OpenAI、Anthropic、Meta、Google、SpaceXAI及Nvidia之代表共同签署。
A spokesperson stated that OpenAI has committed to bringing in external evaluators and to signing contracts with third-party safety assessment organizations.
一名發言人表示,OpenAI已承諾引進外部評估人員,並與第三方安全評估機構簽訂合約。
The spokesperson directed the Associated Press to a company post, which stated that these researchers were dismissed for violating "clear policies on handling sensitive information."
該發言人請美聯社查閱公司的一則貼文,該貼文稱這些研究人員因違反「處理敏感資訊的明確政策」而被解僱。
A spokesperson for Anthropic stated that the company is bringing in independent evaluators and is in discussions with industry groups about collaboration.
Anthropic的一名發言人表示,該公司正引入獨立評估員,並與產業團體商討合作。
Last Friday, Anthropic voluntarily disclosed that its agent had submitted a false tip to the Philadelphia police website regarding an unsolved homicide case.
Anthropic上週五自願披露,其代理曾向費城警方網站提交一則關於一起未破凶殺案的不實線報。
During the administration of Democratic President Joe Biden, leading artificial intelligence companies also voluntarily pledged to promote the "safe, secure, and transparent development" of their models, incorporating specific testing guidelines and transparency reports.
在民主黨總統喬·拜登執政期間,領先的人工智慧公司亦自願承諾,推動其模型「安全、可靠且透明的發展」,並納入具體的測試指引與透明度報告。
However, Sarah Myers West, a former senior AI adviser at the Federal Trade Commission, pointed out that the current agreements lack these elements.
然而,前聯邦貿易委員會資深AI顧問莎拉·邁爾斯·韋斯特指出,現行協議缺乏這些內容。
"On the whole, this self-regulatory approach has consistently done only one thing, and that is fail.
「總體而言,這種自我監管做法始終只做了一件事,那就是失敗。
It has failed in any meaningful way to ensure that these companies even honestly manage their own affairs, let alone ensure that they comply with existing laws."
它未能以任何有意義的方式確保這些公司甚至誠實地管好自己的事,更不用說確保它們遵守現行法律。」
said Myers West, who currently serves as co-executive director of the New York-based nonprofit AI Now.
現任紐約非營利組織AI Now聯合執行董事的邁爾斯·韋斯特說。
"All they have gotten from this administration is leniency."
「它們從這個政府得到的,只有寬容。」
After having lunch with Trump last month, the FTC said it had launched an investigation into the risks that the technology of OpenAI, Anthropic and other companies could pose to consumers.
在上個月與川普共進午餐後,FTC表示已就OpenAI、Anthropic及其他公司之技術可能對消費者構成的風險展開調查。
The FTC did not immediately respond to a request for comment.
FTC未立即回應置評請求。
Trump recently appointed Director of National Intelligence Jay Clayton and Federal Trade Commission Chairman Andrew Ferguson to join his newly established "Super Intelligence Task Force."
川普近日任命國家情報總監傑伊·克萊頓與聯邦貿易委員會主席安德魯·弗格森,加入其新成立的「超級智慧部隊」。
The president stated that this move would allow the United States to maintain its dominant position in the development of artificial intelligence.
總統表示,此舉將使美國在人工智慧發展中維持主導地位。
In a post on Thursday, he said the White House would regard anyone who does not refer to AI using his preferred wording as an "enemy!"
他週四發文稱,白宮將把任何不以他偏好措辭指稱AI的人視為「敵人!」
Graham Steele, a former Assistant Secretary at the US Treasury, has stated that even if AI companies set up internal assessment teams or bring in external evaluators, such arrangements may still leave their work vulnerable to undue pressure.
前美國財政部助理部長格雷厄姆·斯蒂爾表示,即便AI公司組建內部評估團隊或延攬外部評估員,此類安排仍可能令其工作易受不當壓力左右。