獨家：Anthropic悄悄設立生物實驗室，加速推動AI製藥計畫

作者：Jeffrey Dastin、Michael ErmanBy Jeffrey Dastin and Michael Erman

舊金山9月18日（路透社）——Anthropic 已悄悄設立一座實驗室，進行實體生物學研究，將其人工智慧抱負推向藥物科學領域。SAN FRANCISCO, Sept 18 (Reuters) - Anthropic has quietly set up a laboratory to do physical biology work as it pushes its artificial-intelligence ambitions into the world of drug science.

在外界對 AI 的恐懼席捲大眾之際，兩名知情人士表示，這家新創公司已在舊金山灣區建立一座濕實驗室，也就是進行實體實驗的場所。Anthropic 曾表示，希望能開發罕見疾病療法；其生物學研究也已超越「in silico」或電腦評估的範疇。At a time when fears of AI are gripping the public, the startup has built a wet lab, or place for physical experiments, in the San Francisco Bay Area, two people familiar with the matter said. Anthropic has said it wants to unlock treatments for rare diseases, and its biology work has gone beyond "in silico" or computer evaluations.

Anthropic 生命科學部門主管艾瑞克・考德勒-亞伯拉姆斯（Eric Kauderer-Abrams）週二接受路透社訪問時，證實這家新創公司設有濕實驗室。In a Reuters interview on Tuesday, Anthropic's head of life sciences, Eric Kauderer-Abrams, confirmed the startup's wet lab.

他說：「我們認為，要做生物學，最終檢驗現在仍然是、也將有一段時間會是在真正的實驗室工作中完成。我們今天確實正在做這件事，而我會形容我們的做法很典型，就像你在多數生技公司會看到的那樣：有一部分在我們自己的設施中進行，有一部分則與外部夥伴合作。」"We believe that to do biology, the final test is still and will be for a while in real lab work," he said. "We absolutely are doing that today, and I would describe our approach as being typical of what you would see in most biotech companies, where there's some amount of that that we're doing in our own facilities and some amount of that that we're working with external partners on."

一名發言人後來澄清，Anthropic 的實驗室並非專門用於藥物發現，但拒絕進一步說明。A spokesperson later clarified that Anthropic's lab is not for drug discovery specifically, declining to elaborate.

一名知情人士表示，此舉代表 Anthropic 朝其巨大抱負邁出一小步：要處理該公司認為製藥業忽視的疾病，並且趕在員工與大眾失去信心之前，證明 AI 值得承受工作被消除、甚至可能危及人類生命的代價。該公司推動的任何藥物計畫能否成功仍不確定；多數藥物都無法通過臨床安全性與療效試驗。The move represents a small step toward Anthropic’s enormous ambition to tackle diseases that it thinks the pharmaceutical industry is neglecting — and to do so before staff and the public lose faith that AI justifies erasing jobs and possibly human life, one of the people said. The success of any drug program it runs is not certain; most drugs fail to pass trials for clinical safety and efficacy.

對 Anthropic 執行長達里歐・阿莫代伊（Dario Amodei）而言，這項努力帶有個人意義。他在近期一篇文章中表示，他的父親因病去世，僅僅幾年後治療方法才問世。考德勒-亞伯拉姆斯也呼應了這種心情。For Anthropic CEO Dario Amodei, the effort is personal. A disease killed his father only a few years before a cure came about, he said in a recent essay. Kauderer-Abrams echoed the sentiment.

考德勒-亞伯拉姆斯說：「公司的使命，是以造福世界的方式開發強大的 AI。到目前為止，我們認為生命科學是實現這一點的最大機會，這也推動著我們所做的一切。」“The mission of the company is to develop powerful AI in such a way that benefits the world,” Kauderer-Abrams said. “By far, we see the biggest opportunity for that in the life sciences, and that is motivating everything that we’re doing.”

在雄心與災難疑慮之間取得平衡BALANCING AMBITION WITH FEARS OF CATASTROPHE

兩名不願具名受訪的人士表示，Anthropic 和其他生技公司一樣，正在擁抱實體自動化。Like other biotech companies, Anthropic is embracing physical automation, said the two people, who spoke on condition of anonymity.

其中一人表示，這家新創公司希望推進其 Claude AI 指揮機器人單元，在有限人為介入下執行科學實驗的能力。不過，Anthropic 發言人表示，公司認為人類監督與參與對安全至關重要。The startup wants to push how its Claude AI can direct robotic units to carry out science experiments with limited human intervention, one of the people said. Still, Anthropic believes that human oversight and involvement are essential for safety, its spokesperson said.

考德勒-亞伯拉姆斯說：「我們才剛開始使用 AI 來自動執行實驗室工作，這個領域有潛力讓許多不同流程出現有意義的加速。」"We're in the very early innings of using AI to automate the execution of lab work," said Kauderer-Abrams, "an area that has the potential to bring about meaningful acceleration in so many different processes."

Anthropic 競相把快速演進的技術應用到生物實驗室，正值這家公司非比尋常的時刻。Anthropic’s race to apply rapidly evolving technology to a biology lab comes at an extraordinary time for the company.

過去兩週，該公司部分研究人員警告，AI 可能導致人類滅絕；Anthropic 也表示，發現其系統可能被用於生物武器開發的例子，但這項說法引來一些質疑。In the past two weeks, some of its researchers warned that AI could lead to human extinction, and Anthropic said it found examples where its systems could have been used toward biological weapons development, which met some skepticism.

阿莫代伊也呼籲放慢開發腳步，稱未來若出現比曾駭入 Hugging Face 的 OpenAI 代理程式更有能力的 AI，可能造成災難性破壞。與此同時，Anthropic 正準備上市，進行規模可能達2兆美元的首次公開募股（IPO），為這類 AI 開發籌集更多資金。Amodei also called for a development slowdown, saying that a future AI more capable than the OpenAI agents that hacked Hugging Face could cause catastrophic damage. At the same time, Anthropic is preparing to go public in a potentially $2 trillion IPO to raise more money for such AI development.

考德勒-亞伯拉姆斯說：「在我們的生命科學工作中，我們一直在平衡這些事情：以我們認為能提升並加速藥物開發的方式讓模型變得更好，同時也要審慎思考並謹慎部署，以降低風險。」"In our life-science work, we're always balancing these things," Kauderer-Abrams said, "making models better in ways that we think will uplift and accelerate developing medicines, and being thoughtful and careful about how we deploy it to limit the risk."

擴大採購PROCUREMENT RAMP-UP

Anthropic 今年6月在舊金山一場活動中，表明有意開始推動藥物計畫。考德勒-亞伯拉姆斯當時在台上表示，Anthropic 將在傳統公司認為財務上不具吸引力的領域進行臨床前工作。Anthropic staked out its intention to start running drug programs, at a San Francisco event in June. On stage there, Kauderer-Abrams said Anthropic would do preclinical work in areas that traditional companies did not find attractive financially.

該公司也推出名為 Claude Science 的軟體；據媒體報導，Anthropic 先前還讓諾華執行長瓦斯・納拉辛漢（Vas Narasimhan）加入董事會，並以約4億美元股票收購新創公司 Coefficient Bio。Anthropic 證實這項收購案，稱其將有助於打造藥物開發工具，但未對交易價格作出回應。The company also launched software called Claude Science, and it earlier added Novartis CEO Vas Narasimhan to its board and bought the startup Coefficient Bio for about $400 million in stock, according to media reports. Anthropic confirmed the acquisition, which will help it build tools for drug development, but it had no comment on the deal price.

考德勒-亞伯拉姆斯告訴路透社，如今 Anthropic 正擴充內部實驗室工作，以追求速度，並取得生物學第一手經驗；不過在更有效率時，該公司也會將工作外包。Now, Anthropic is bulking up its in-house lab efforts for speed and for firsthand experience in biology, though it outsources work when efficient, Kauderer-Abrams told Reuters.

他說：「有些事情我們親自做可以快得多」，目標是「以盡可能大的規模運作」。“There’s some things that we can do much faster in our own hands,” he said, with the goal “to operate at the largest possible scale.”

Anthropic 近期在 LinkedIn 徵才貼文中表示，公司希望聘請一名能擴大採購與其他營運工作的主管。Anthropic 透過 Jobera 招募的另一個職缺，尋找精通「蛋白質與核酸表徵」以及其他生化工作的專家。Anthropic 第三則職缺寫道：「我們的目標是讓生命科學的進展加快一個數量級。」Anthropic has looked to hire a leader who could ramp up procurement and other operations, it said in a recent LinkedIn job post. Another Anthropic role solicited via Jobera sought an expert on “protein and nucleic acid characterization,” among other biochemical work. A third Anthropic job listing read: “Our goal is to speed up progress in the life sciences by an order of magnitude.”

考德勒-亞伯拉姆斯表示，以人員與資源來看，生命科學已是 Anthropic 最大的投資領域之一。Life sciences already represents one of Anthropic’s biggest investment areas by headcount and resources, Kauderer-Abrams said.

鎖定「無藥可治」疾病TARGETING 'UNDRUGGABLE' CONDITIONS

他的看法是，AI 可能加速開發過去被認為「無藥可治」、也就是太難治療的疾病。舉例來說，這項技術可加快雙特異性與三特異性抗體的發現；這些複雜分子可被導向目標蛋白質或細胞上的多個位置，或同時攻擊多個標的。His view is that AI could fast-track development for conditions previously thought to be "undruggable" — too difficult to treat. The technology could speed discovery, for instance, of bispecific and trispecific antibodies, complex molecules that can be directed at multiple points on a targeted protein or cell, or hit multiple targets.

Anthropic 正在處理哪些具體疾病，以及進展如何，目前仍不清楚。一旦辨識出一個分子，仍可能需要多年時間透過人體試驗測試，才將藥物推向市場；這是一項 Anthropic 尚未承擔的挑戰。競爭對手 Alphabet 旗下藥物發現部門 Isomorphic Labs 已努力多年，希望推進到臨床階段，且先前將這項目標延後到2026年底。The precise diseases Anthropic is addressing and its progress remain unclear. Once a molecule is identified, it can take years to bring a drug to market as it is tested in human trials, a challenge that Anthropic has not taken on. Isomorphic Labs, the drug discovery arm of rival Alphabet, has worked for years to get to the clinic and earlier delayed that goal to the end of 2026.

在推動實驗室與其他產業自動化的努力中，Anthropic 於8月推出 Model Hardware Standard，協助 AI 操作設備。學界先前發布的 PyLabRobot 與 Anthropic 無關，處理的是相關但不同的自動化需求。In its efforts at lab and other industry automation, Anthropic launched the Model Hardware Standard in August, which helps AI operate equipment. A release out of academia called PyLabRobot, unrelated to Anthropic, earlier addressed related though different automation needs.

Anthropic 正把 Model Hardware Standard 這類工作，以及更廣泛的 AI 服務與工具，提供給各式各樣頂尖製藥公司，例如羅氏（Roche）旗下的 Genentech、必治妥施貴寶（Bristol Myers Squibb）與諾和諾德（Novo Nordisk）。Anthropic is providing work like the Model Hardware Standard and broader AI services and tools to all manner of top pharmaceutical companies, such as Roche’s Genentech, Bristol Myers Squibb and Novo Nordisk.

與此同時，兩名熟悉其工作的人士表示，Anthropic 正面臨信任問題。即使這家 AI 新創公司已把客戶資料隔離，避免被查看，其客戶仍可能擔心 Anthropic 會從彼此競爭的藥物計畫中學習。At the same time, Anthropic is navigating a trust problem, the two people familiar with its efforts said. Its customers may worry that Anthropic will learn from their competing drug programs, even as the AI startup walls off their data from view.

考德勒-亞伯拉姆斯表示，部分出於競爭疑慮，Anthropic 已劃下界線：目前不進行臨床試驗，並聚焦於產業不會處理的需求。Kauderer-Abrams said that Anthropic, due partly to competition concerns, has set a boundary that it is not running clinical trials for now and is focusing on needs that industry would not address.

他說：「我們並不是在與以將藥物推向市場為業的製藥和生技公司競爭。」"We’re not competing with pharma and biotech companies that make their business in bringing drugs to market," he said.

（舊金山 Jeffrey Dastin、紐約 Michael Erman 報導；Caroline Humer、Kenneth Li、Matthew Lewis 編輯）(Reporting by Jeffrey Dastin in San Francisco and Michael Erman in New York; Editing by Caroline Humer, Kenneth Li and Matthew Lewis)