本转录稿为2026年10月11日播出的《与玛格丽特·布伦南面对面》节目中，帕兰蒂尔（Palantir）首席技术官希亚姆·桑卡尔（Shyam Sankar）专访的完整文字实录，部分内容已在节目中播出。
2026-10-11T10:31:11-0400 / 哥伦比亚广播公司新闻
以下是与帕兰蒂尔首席技术官希亚姆·桑卡尔的完整访谈实录，部分内容已在2026年10月11日《与玛格丽特·布伦南面对面》节目中播出。
玛格丽特·布伦南： 现在我们邀请到了帕兰蒂尔的首席技术官希亚姆·桑卡尔。欢迎来到《面对全国》节目。
希亚姆·桑卡尔： 感谢邀请我来，玛格丽特。
玛格丽特·布伦南： 很高兴你能来到演播室现场。我们近期一直在大量讨论科技、人工智能以及公众对其的看法，这正是我想从你这里展开的话题。我们的民调显示，目前这一领域存在问题。哥伦比亚广播公司的民调显示，50%的受访者认为人工智能带来的问题多于其解决的问题，29%的人认为其利大于弊，还有21%的人完全不知情。45%的美国人希望政府限制人工智能的使用和开发。你的立场是目前不应设置任何监管护栏吗？
希亚姆·桑卡尔： 不，完全不是。这不是我的立场。我认为——科技行业肩负着绝对的责任，要向美国民众证明人工智能确实能带来切实好处，而不是那种虚无缥缈的好处，比如帮我安排行程，而是能为美国民众带来就业增长和繁荣的实实在在的益处。我认为当前讨论中一个非常复杂的点在于，有一股无形的力量在主导人工智能安全的叙事，它实际上就是这个融合了末日邪教和救世主情结的运动，名叫有效利他主义。大多数美国人最熟悉的有效利他主义者是萨姆·班克曼-弗里德，而有效利他主义者几乎就像一个宗教团体，他们从21世纪初就一直在思考人工智能带来的末日场景。
玛格丽特·布伦南： 他——他现在入狱了，他是那位加密货币投资者——
希亚姆·桑卡尔： 没错。
玛格丽特·布伦南： 他曾欺骗投资者。但这听起来非常学术化。你为什么如此关注这件事？
希亚姆·桑卡尔： 因为美国的人工智能研究人员大概只有2000人左右，其中绝大多数都信奉这种围绕有效利他主义的世界观。如果你听到像雅各布·科克森这样的人说人工智能有10%的概率会终结人类，而你相信——
玛格丽特·布伦南： 这位研究人员最近发出了这样的警告。
希亚姆·桑卡尔： 是的，没错。如果你认为他是基于专业背景发表的技术层面的科学批评，那么美国民众完全有理由对此感到极度恐慌和担忧。但如果你明白这本质上是一种神学层面的批判，你就可以将其与现实中正在发生的经验事实进行权衡。我认为——帕斯卡有句名言，每个人心中都有一个上帝的空缺，这句话道出了深刻的真相。要明白，硅谷许多开发这项技术的人已经用通用人工智能这个“机器上帝”填补了这个空缺。关于这一点有很多有趣的故事，比如，很多人工智能研究人员开始抽烟，以此向其他研究人员表明他们对人工智能的信仰有多坚定。我如此坚信——（对话被打断）
玛格丽特·布伦南： 什么？这完全说不通。你是怎么认为的——
希亚姆·桑卡尔： 我如此坚信人工智能能够治愈癌症，所以我要向你们展示我已经押上了全部筹码。目前存在一种群体性的心理偏执和社会妄想，我认为这正在阻碍我们真正理解人工智能的发展现状。人工智能正在工业领域与美国劳动者产生怎样的结合？这样做的结果是什么？而那些末日论者的预测基本上都错了。他们预言的所有事情都没有成真：他们说ChatGPT 2.0发布过于危险，他们说按照当前的发展节奏，我们早就该面临大规模失业和岗位替代了，但经济学家们发现，实际上我们在一些此前无法预见的领域实现了就业增长。所以科技行业肩负着双重责任：一是要向美国民众清晰地表明，人工智能将为国家带来经济繁荣和安全；二是要切实承担起工程设计的责任。你不能逃避“设计安全产品是你的职责”这一理念，不能对此含糊其辞。如果你看看末日论者的言论，会发现它们是相互矛盾的。一种说法是这项技术会终结人类。
玛格丽特·布伦南： 没错。
希亚姆·桑卡尔： 与此同时，他们又说“但我相信它有一定概率创造乌托邦”。乍一看这似乎自相矛盾，但它们有一个共同点：作为人类，你没有任何主动权。机器上帝将决定未来，或许只有少数人能够试图避免可怕的结局。我认为这掩盖了一个更深刻的现实：我们人类拥有巨大的主动权，可以决定如何利用这项技术实现美国民众期望的成果。
玛格丽特·布伦南： 那我们来谈谈这个问题。另一种反驳观点会说，那些反对设置监管护栏的人纯粹是出于商业目的，他们只想继续销售芯片，或者只想抢占市场。所以这是对你所说观点的反驳，也就是你反对设置监管护栏是出于意识形态立场。
希亚姆·桑卡尔： 或许我们可以先退一步，谁是反对设置监管护栏的一方？
玛格丽特·布伦南： 好吧，我想深入探讨具体细节，因为就连“监管护栏”的定义都非常复杂，目前甚至没有达成共识。但新任国家情报总监兼人工智能沙皇杰伊·克莱顿本周表示，工作重点是管控和监测那些有可能造成伤害的技术。在你看来，这具体意味着什么？听起来他自己也还不确定。
希亚姆·桑卡尔： 我认为我们应该明确，我们应当监管这项技术的应用场景。如果你将其用于医药领域，就像美国食品药品监督管理局（FDA）的监管逻辑一样。作为一个共和国，我们拥有250年丰富的法律和历史经验来管理责任问题。我不想把这个比喻说得太简单，但举个例子，如果我买了一辆有缺陷的汽车，那是汽车制造商的问题；但如果我开着那辆车撞了邻居的房子，那就是我的问题。而现在这种叙事混淆了概念，声称这项技术是完全不同性质的事物。我不这么认为，我认为它只是程度上的差异。
玛格丽特·布伦南： 好吧，如果人类无法掌控它，那就是性质不同的事物了。
希亚姆·桑卡尔： 不，人类可以掌控它。即使我们看那些围绕Hugging Face平台的耸人听闻的报道，那些模型也没有做出任何超出指令的行为。它们——最初的问题出在人为失误。它们没有处于沙箱环境中，而是接入了互联网，这是一个糟糕的网络安全配置。从这个角度来看，它们被要求在特定基准测试中取得成功，并且没有得到实现目标的约束条件。于是它们说：“要在基准测试中取得成功，最好的办法就是获取考试答案，然后照做。”
玛格丽特·布伦南： 那么，针对这种情况，谁应该采取行动来——
希亚姆·桑卡尔： ——我认为这是一款有缺陷的产品。就像——
玛格丽特·布伦南： ——应该由司法部对程序员提起诉讼吗？比如，谁应该为这项技术没有局限在沙箱环境中承担责任——
希亚姆·桑卡尔： ——当然是人工智能公司。在那个具体案例中，责任方是人工智能公司。如果另一家公司获得授权使用这项技术开发应用，产品按预期运行，但你误用了产品，那么责任就在你身上。
玛格丽特·布伦南： 所以我的理解是，你认为现有的法律足以应对技术的发展？
希亚姆·桑卡尔： 我认为现有的法律是一个巨大的起点。我认为现行法律已经涵盖了当前的诸多考量。或许我们需要在边缘领域补充一些条款？当然，但我认为我们可以在实践中逐步完善。
玛格丽特·布伦南： 那我们来谈谈其中一些具体问题。你曾陪同总统会见近期被召集的人工智能领域领军人物，当时总统要求他们签署一项承诺，承诺实施强有力的内部管控、与外部审计机构合作，并设立委员会评估审计报告。谷歌、Anthropic、OpenAI、英伟达、埃隆·马斯克都签署了这项承诺。而你们公司帕兰蒂尔没有签署。为什么？
希亚姆·桑卡尔： 嗯，这些协议是针对那些真正开发人工智能模型的公司的，而我们公司并不开发模型，所以——
玛格丽特·布伦南： 但你们会使用这些模型并搭建平台。所以你不认为你们需要设置相关约束？
希亚姆·桑卡尔： 哦，我们确实有这些约束。事实上，这正是我们整个业务的一部分，即如何在特定应用场景中安全部署这些模型。我们只是没有参与这项协议——我们没有被邀请加入该协议。
玛格丽特·布伦南： 但你出席了那场会议。
希亚姆·桑卡尔： 我出席了。
玛格丽特·布伦南： 那么，当时的谈话是怎样的？
希亚姆·桑卡尔： 谈话的核心是，美国民众期望人工智能能够创造经济繁荣，并且是在安全的前提下实现这一目标，而我们作为企业，有明确的义务切实兑现这些承诺，我们可以依靠现有法律来处理相关问题，我们都应该认真对待这件事。那是一场非常有成效的对话。
玛格丽特·布伦南： 因为听起来白宫当时是在说需要采取更多措施，而现在新任人工智能沙皇杰伊·克莱顿本周将前往硅谷。他表示已启动为期120天的人工智能政策审查。你希望从中看到什么成果？还是说这又是另一个不会产生任何实际结果的蓝丝带委员会？
希亚姆·桑卡尔： ——不，我认为这次审查会带来很多成果，核心是聚焦和投入精力，利用现有机构和监管体系确保我们能掌控局势，并在需要的地方提供支持。我认为其中一个挑战，回到我之前关于有效利他主义的评论——目前负责监管的所谓独立研究组织实际上并不独立。它们与相关实验室存在利益冲突，其成员都持有单一的有效利他主义世界观，一直在寻找末日场景。你会看到评估人员本质上是举着一把上膛的枪对准目标，然后说“我不敢相信枪真的响了”。而他们真正应该问的是：“我们设置了哪些监管护栏？这些产品是否符合我们的设计规格？存在哪些被滥用的可能性？”
玛格丽特·布伦南： 所以我曾与金融行业的人士交谈过，他们说我们确实需要一些监管，但美国证券交易委员会（SEC）、商品期货交易委员会（CFTC）等机构已经在监管相关业务了。你不认为在现有机构框架内，可以设立一个专门监管技术的实体吗？
希亚姆·桑卡尔： 我认为我们需要建立一个事件响应和报告机制，在产品发布前就提高测试过程中的可见性和透明度，让政府能够了解企业如何落实现有法律要求。
玛格丽特·布伦南： 谁来负责这项工作，或者说谁可以负责？
希亚姆·桑卡尔： 我会 defer 给克莱顿局长，但我认为这是他们可能会推进的方向之一。
玛格丽特·布伦南： 他得弄清楚这个问题。我们看看上一届政府——拜登总统和习近平主席曾达成一项非常有限的协议，但至少在一个原则上达成了共识，即人类而非人工智能应该保留使用核武器的决策权。这是否合理——
希亚姆·桑卡尔： 嗯，这似乎是显而易见的。
玛格丽特·布伦南： 首先，他们不得不坐下来讨论这件事，这本身就令人不安，但这不应该在国防领域的其他方面推广吗？你同意吗？这是不是其中一个需要——
希亚姆·桑卡尔： 嗯，当前的政策分为两部分：一部分是军事 doctrine，即人类始终对每一项决策负责；另一部分是，在涉及致命武力的决策过程中，人类必须参与其中。
玛格丽特·布伦南： 但在你认为应该设置更多监管护栏的领域，国防领域是不是其中之一？生物技术领域，我听其他人也这么说过。
希亚姆·桑卡尔： 所有领域都需要——都需要考虑应用安全。
玛格丽特·布伦南： 但这仍然存在一个问题：谁来负责，以及如何负责？
希亚姆·桑卡尔： 嗯，我认为国防部和国会应该负责国防领域的安全监管，交通部负责交通领域，FDA负责医药领域。
玛格丽特·布伦南： 所以，预计这方面可能会有所行动。因为你知道，科技行业的人会说，我们需要全速前进以与中国竞争。中国正在推动人工智能全面融入其所有产业，但他们有能力放缓发展速度以跟上监管步伐，他们之前就这么做过。所以你认为美国也应该考虑在某些领域踩刹车，先放缓发展速度，直到明确监管护栏的方向？
希亚姆·桑卡尔： 你知道，美国拥有这套美丽而混乱的流程，但尽管看起来令人沮丧，我认为它正在发挥作用。我们现在已经有了尚未发布的前沿模型，因为人们意识到我们可能还没准备好发布它们。这在一年前是不可想象的。随着人们开始思考实验室、政府应该如何作为，我认为这正是我们混乱的流程和各方讨论带来的结果，我们正在逐步改进实施方式，发展速度自然会随之改变。
玛格丽特·布伦南： 我们来谈谈国防部。他们最近停止使用Anthropic的产品，原因是该公司坚持要求对军事用途的人工智能设置保护措施。我知道你了解这个案例。Anthropic的Claude曾是Maven平台的一部分，而帕兰蒂尔也使用该平台。这一决定会产生多大影响？谁会取代他们的位置？
希亚姆·桑卡尔： 嗯，目前国防部拥有广泛的模型供应商选择，从XAI和OpenAI到英伟达的Nemotron等开源模型，所以我不认为——我不认为这会对运营产生影响。我认为这背后有一个更广泛的声明，实际上推动了我们商业业务的增长，即企业正在意识到，这同样适用于政府：你不会只选择一家模型供应商。实际上，作为一个机构，你的主权和能动性取决于你能否掌控、决定在何种场景下使用哪些模型，能否切换模型，能否自主控制模型权重——在你需要感受到这种控制权的地方。这是你的核心竞争力，是让你的机构与众不同的关键。
玛格丽特·布伦南： 所以你还不确定谁会取代他们，但你不认为这一决定会产生重大影响。这在硅谷和国防领域无疑是个大新闻——
希亚姆·桑卡尔： 已经有替代方案了，已经找到了替代方案。开发人工智能应用的用户可以自主选择他们使用的模型。可用的模型套件包括谷歌、XAI、OpenAI提供的产品。你刚才的第二个问题是什么来着，抱歉？
玛格丽特·布伦南： 是的，我刚才问的是影响。听起来你认为这不会——不会造成太大影响。好吧，我可以继续了。
希亚姆·桑卡尔： 没有影响。
玛格丽特·布伦南： 你最近曾表示，伊朗战争是第一场人工智能发挥核心作用的重大冲突，尤其是在目标定位方面。你认为这其中存在风险吗？
希亚姆·桑卡尔： 嗯，你知道，“目标定位”这个词有些耸人听闻。实际上，目标定位是一个流程。如果你想想军事行动，以前，一天执行1000次打击需要50人花费两周时间来规划，而现在一个人花4小时就能完成，这能让你在战场上相对于对手获得显著优势。所以这并不是人们立刻联想到的那种“人工智能挑选目标”的情况，实际并非如此。它真正的作用是，当你明确了想要达成的效果——比如我想摧毁敌人的防空系统。好的，哪些目标与此相关？好的，接下来我如何将它们纳入流程，确保这些目标不在打击黑名单上？是否有任何目标距离民用基础设施过近？在时间紧迫的情况下，如何完成所有这些极易出现人为失误的环节？以一种可重复、高精度的方式完成这些工作，从而提升系统的处理能力。
玛格丽特·布伦南： 嗯，这涉及到一个更广泛的争论：削减那些负责减轻打击对平民影响的人员岗位。这是赫格斯塞特部长做出的五角大楼决定。你刚才谈到了数千次打击，伊朗的目标定位行动中有很多成功的案例。但你也知道，米纳布地区发生了一起可怕的事件，造成至少123名学童死亡。彭博社报道称，一些人员过度依赖帕兰蒂尔公司开发的、采用人工智能技术的Maven系统。你是否确定当时发生了什么？
希亚姆·桑卡尔： 嗯，我们应该等待最终报告。但我可以说的是，相关报道并不准确。
玛格丽特·布伦南： 你是说彭博社的报道不准确？
希亚姆·桑卡尔： 是的，报道失实，实际原因很可能是人为失误。而且——在那之后不久，我们就开发了更多的人工智能代理来进行更多检查。这是我们无法掌控的流程的一部分，但为了应对当下的紧迫性，我们开发了额外的人工智能工具，在我们无法掌控的流程之外进行更多复核，以确保类似事件不再发生。
玛格丽特·布伦南： 所以你是在暗示，是情报过时或者人为失误将信息输入了计算机系统。你是这个意思吗？
希亚姆·桑卡尔： 嗯，我们应该等待最终报告，但我认为我们会发现——
玛格丽特·布伦南： 你知道报告何时发布吗？
希亚姆·桑卡尔： 我不知道。
玛格丽特·布伦南： 因为最初的报道已经被广泛传播，包括本媒体在内。你认为尽快公布相关信息是否有坏处，还是说这对你所说的想要做出的改进有所帮助？
希亚姆·桑卡尔： 嗯，我会 defer 给国防部来决定——我无法控制发布时间。但没错，我希望真相能够公之于众。
玛格丽特·布伦南： 但你认为那是人为失误。你是这个意思吗，而非彭博社报道所说的，他们过度依赖人工智能。
希亚姆·桑卡尔： 彭博社的报道是错误的。没错。
玛格丽特·布伦南： 有120名众议院民主党人就这起事件以及人工智能在打击行动中的应用，包括Maven系统，致信赫格斯塞特部长。你是否打算向国会简要介绍相关情况或分享信息？
希亚姆·桑卡尔： 我们已经向国会进行过简报，而且——他们也清楚Maven系统的能力。
玛格丽特·布伦南： 但就这起打击行动以及哪里出了问题，显然确实发生了失误。
希亚姆·桑卡尔： 我会 defer 给国防部来解释。那——不是我们的职责范围。
玛格丽特·布伦南： 但帕兰蒂尔与美国政府的联系如此紧密。我的意思是，你们目前拥有数十亿美元的政府合同。你是否认为应该向国会通报更多关于这项技术的益处和风险的信息？
希亚姆·桑卡尔： 遗憾的是，我只是一名技术专家。在这些政策问题上，我会 defer 给相关人士。
玛格丽特·布伦南： 好吧，在节目结束前，我想快速问一下你参与的现代化建设工作。我了解到你是陆军预备役部队陆军执行创新团的中校。
希亚姆·桑卡尔： 没错。
玛格丽特·布伦南： 这是一个新设立的职位。
希亚姆·桑卡尔： 是的。
玛格丽特·布伦南： 还有来自Meta和OpenAI的高管也加入了该团队。那么你在那里负责什么工作？
希亚姆·桑卡尔： 我的主要职责是帮助陆军G-1部门负责人力资源工作。也就是制定战略，帮助——陆军拥有大量人才，我们称之为“绿衣人”，也就是真正懂代码的士兵，他们——他们非常优秀。我们该如何任用他们？你知道，他们不符合工业时代的模式，不符合拿破仑式的军衔体系。他们的技能与军衔无关。你可能会遇到一个E-4军衔的士兵，却是一名非常优秀的程序员，但在陆军内部并没有明确的职业发展路径。陆军不知道该如何管理这类人才。那么，我们该如何识别这些优秀人才？如何保护他们？如何将他们安排到能充分发挥技能的项目中，将他们转化为军队的宝贵资产？
玛格丽特·布伦南： 所以这就是你在这个团队中的职责？
希亚姆·桑卡尔： 没错。
玛格丽特·布伦南： 那么对于那些认为这些大型企业可能存在利益冲突的人，因为他们也在与美国政府开展业务。你对此有何回应？
希亚姆·桑卡尔： 嗯，国防部的律师团队会进行严格管理。我们只能参与那些不存在利益冲突的项目。
玛格丽特·布伦南： 那么你认为——
希亚姆·桑卡尔： 我想更广泛地说，能够为美国军队服务是我最大的荣幸。你知道，作为一名移民到这个国家的人，这是——我已故的父亲一直希望我能进入军事院校。人生的轨迹就是这样。我希望在44岁的年纪，通过这个身份为国家做出的贡献，能比我24岁时更多，这让我深感自豪。
玛格丽特·布伦南： 希亚姆，感谢你今天抽出时间接受采访。
希亚姆·桑卡尔： 谢谢。
玛格丽特·布伦南： 我们稍后回来。
Transcript: Shyam Sankar, chief technology officer of Palantir, on “Face the Nation with Margaret Brennan,” Oct. 11, 2026
2026-10-11T10:31:11-0400 / CBS News
The following is the full transcript of the interview with Shyam Sankar, chief technology officer of Palantir, a portion of which aired on “Face the Nation with Margaret Brennan” on Oct. 11, 2026.
*
MARGARET BRENNAN: And we’re joined now by the Chief Technology Officer of Palantir, Shyam Sankar. Welcome to Face the Nation.
SHYAM SANKAR: Thank you for having me, Margaret.
MARGARET BRENNAN: It’s good to have you here in person. We’ve been talking a lot about technology, artificial intelligence, and public perception of it right now, and that’s where I want to start with you because our polling shows there’s a problem here. 50% of those polled by CBS say AI will create more problems than it solves. 29% say it’ll solve more than it creates, and 21% just really don’t know. Like 45% of Americans want the government to restrict use and development of AI. Is your position there should not be any guardrails at this moment?
SHYAM SANKAR: No, not at all. That’s not my position. I think the- the burden is absolutely on the technology industry to show the American people that there are real benefits to AI and not kind of ephemeral sort of benefits like it can help me schedule travel, but actually meaningful benefits that result in job growth and prosperity for the American people. I think one of the things that’s really complicated in the current discussion, is that there’s an unseen force driving the narrative around AI safety, and it’s- it’s really this millenarian doomsday cult meets messianic complex called effective altruism. The most famous effective altruist that most Americans know is Sam Bankman-Fried, and the effective altruists are, it’s almost like a religion that they’ve been thinking about doomsday scenarios from AI since the early 2000s.
MARGARET BRENNAN: He was–he is jailed. He was the crypto investor—
SHYAM SANKAR: Correct.
MARGARET BRENNAN: Who defrauded investors. But this sounds very academic. Like, what makes you so focused on this?
SHYAM SANKAR: Because there’s only really roughly 2,000 AI researchers in America, and the vast majority of them believe in this sort of worldview around effective altruism. And if you hear someone like Jacob Coxon telling you that AI has a 10% chance of ending humanity, and you believe—
MARGARET BRENNAN: A researcher who had this warning recently.
SHYAM SANKAR: Yes, correct. And you believe he’s speaking as a technically informed individual who does this work, it’s a scientific critique. I think the American people would be right to be quite panicked and worried about what’s going on here. But if you understand that it’s actually a theological critique, then you can- you can kind of weigh it against the empirical facts of what is actually happening out there. You know, I take—I think there’s a profound truth to Pascal’s saying that every human has a god-shaped hole in their heart, and it’s important to understand that many of these folks in Silicon Valley who are building this technology have filled that hole with the machine God of AGI. And I mean, there’s so many interesting stories around this. Like, for example, so many AI researchers have started smoking as a way of signaling to other researchers how much they believe in AI. I believe so much— (CROSSTALK)
MARGARET BRENNAN: What? That makes no sense. What do you think—
SHYAM SANKAR: I believe so much in AI that it’s going to cure cancer. So I will—I’m showing you my chips are all in. So there is kind of this mass psychosis, social delusion around it, and I think it’s getting in the way of actually what is happening with AI. Where is it meeting the American worker in industry? What are the results of doing that, and the doomsdayers have basically been wrong. Everything they predicted, they- they told us that ChatGPT 2.0 was too dangerous to release. They told us at this point in the timeline, we’d already have mass unemployment and job displacement, and the economists see that actually we’re having job growth in areas that we wouldn’t have predicted doing exactly that. So the technology industry has two burdens. One is to show very clearly to the American people that- that AI is going to result in economic prosperity and security for the nation. And two, to actually do and absorb the burden of engineering. You cannot abdicate this idea that designing safe products is your job. You can’t hand wave around it. If you look at the kind of narratives that come out of the doomsdayers, they seem like two contradictory narratives. One narrative is this technology is going to end humanity.
MARGARET BRENNAN: Right.
SHYAM SANKAR: And simultaneously, they say, “But I believe it has some chance of creating utopia.” So at first blush, this seems contradictory, but it has one thing in common, which is that you, as a human, have no agency. The machine god is going to decide the future, and only perhaps a small number of us can try to avoid the horrible outcome. And I think it obfuscates the more profound reality that we have a huge amount of human agency over how we’re going to use this technology to deliver the outcomes the American people expect.
MARGARET BRENNAN: Well let’s talk about that because the other part of this argument would be, and no one who wants- the people who don’t want guardrails are simply in it for commercial reasons. They want to keep selling chips, or they just want access to market. So that would be the counter argument to the idea that you know what you’re saying, which is an ideological opposition to guardrails.
SHYAM SANKAR: Maybe we could just back up and who’s on the side of no guardrails?
MARGARET BRENNAN: Well, when- I want to get into the specifics here, because even just defining what guardrails is seems really complicated, and there’s not even common agreement on that right now. But Jay Clayton, who is the new DNI and the new AI czar, said this week the focus is on control and monitoring of technologies with the potential to do harm. So, what does that look like to you? Because it sounds like he’s not even sure yet.
SHYAM SANKAR: Well, I think there’s a very clear view that we should be regulating the application of the technology. So if you’re going to use this in medicines, you know the FDA- As a republic, we have 250 years of rich laws and history on how to manage liability. I don’t want to trivialize the metaphor, but you know, if I buy a car and the car is defective, that’s the car maker’s problem. If I then use that car to go drive into my neighbor’s house, that’s my problem. And you know, this sort of obfuscation. The narrative is this technology is a total difference of kind. I don’t think so. I think it’s a difference of degree.
MARGARET BRENNAN: Well, in that humans aren’t in control of it, it would be a different kind.
SHYAM SANKAR: No, humans are in control of it. Even if we look at the kind of salacious reporting around Hugging Face, the models did not do anything they were not instructed to do. You know, they were- the- the real first, you have human error. They- they were not in a sandbox. They had access to the internet. It was a horrible cybersecurity configuration. Then, from that perspective, they were asked to- to- succeed at a specific benchmark, and they were not given constraints on how to do that. And they said, “Oh, the best way to succeed at the benchmark is to get the answers to the test, and to do so.
MARGARET BRENNAN: So, for that, who should take action to–
SHYAM SANKAR: –I think it’s a defective product. Like, if–
MARGARET BRENNAN: –Is it up to the Justice Department to then bring suit, like on the programmers, like who’s liable for the technology not, you know, staying within the sandbox–
SHYAM SANKAR: –The AI companies of course. In that, in that specific case, it would be the AI companies. If another company had licensed that technology to use in another in an application, and the product worked as expected, but you misused the product, then it would be you that are responsible for that.
MARGARET BRENNAN: So you believe that, am I understanding you, the existing laws on the books should be able to handle technology as it develops?
SHYAM SANKAR: I think it is a massive running head start. I think the laws on the books speak to a lot of the current considerations. Might we need a few more things on the margin? Sure, but I think we’ll figure that out as we go through it.
MARGARET BRENNAN: So let’s talk about some of those things because you were with the president when he summoned some of these AI leaders recently, and he asked them all to sign a pledge to implement robust internal controls, partner with an external auditor and establish a committee to evaluate reports from auditors. That was Google, Anthropic, OpenAI, Nvidia, Elon Musk. They all signed it. Palantir, your company didn’t. Why?
SHYAM SANKAR: Well, those accords were for companies that actually make AI models, which we do not, so–
MARGARET BRENNAN: But you help to use them and build out platforms. Like you don’t see a reason for constraints there.
SHYAM SANKAR: Oh, we- we do have those constraints. In fact, that- part of that’s our entire business, is how do you safely deploy these models in specific use cases. We’re just not a party that- we weren’t asked to be a part of the accord.
MARGARET BRENNAN: But you were there at that meeting.
SHYAM SANKAR: I was.
MARGARET BRENNAN: So, what happened? Like, what was the conversation?
SHYAM SANKAR: The conversation was really an assertion that the American people expect AI to generate economic prosperity and to do so securely, and that we as companies have an affirmative obligation to deliver on exactly those things, and that we have existing laws to look at this, and that we should be all taking this quite seriously, very productive conversation.
MARGARET BRENNAN: Because it sounded like the White House was saying something more should be done, and now you have this new AI Czar Jay Clayton going out to Silicon Valley this week. He said he’s launched this 120-day review of AI policy. What do you hope comes from that, or is this another one of these like blue ribbon commissions that doesn’t produce any–
SHYAM SANKAR: –No, I think a lot will come up to it, which is really focus and energy, and using the existing agencies and regulatory apparatus to make sure we’re on top of this stuff and providing support where it’s needed. I think one of the challenges, going back to my comments on effective altruism is, today, the so-called independent research organizations that are overseeing this are not really independent. They have conflicts with the labs. They are made up of people with a singular worldview around effective altruism, looking for the doomsday scenarios. You know, you have evaluators who are essentially pointing a loaded gun at a target, saying, “I can’t believe the gun worked.” When they should really be saying, “What guardrails do we have in place? You know- are you know are these products working to the specifications we designed, and what- what are the opportunities for misuse?
MARGARET BRENNAN: So I’ve spoken to people, for example, in the financial industry who say we did need some regulation here, but there’s the SEC, there is you know the CFTC, they oversee certain operations here. Don’t you think within the current institutions there could be some sort of entity that oversees technology?
SHYAM SANKAR: I think there’s going to need to be an incident response and reporting regime that creates more visibility and transparency into what you’re learning in the testing, well ahead of product launches that give visibility and transparency to the government of how people are doing and implementing exactly the existing laws.
MARGARET BRENNAN: Who does that, or who could do that?
SHYAM SANKAR: I will defer to Director Clayton on that, but I think that’s one of the things that I- I suspect they will be driving towards.
MARGARET BRENNAN: He’s got to figure that out. Well, we look at the last administration- President Biden and Xi Jinping came to like, a very limited agreement, but at least on one principle, which was humans, not AI, should retain control over decisions to use nuclear weapons. Would it be sensible–
SHYAM SANKAR: Well, that seems obvious.
MARGARET BRENNAN: Well, that seems terrifying, first of all, that they had to come to a conversation about that, but shouldn’t that be expanded elsewhere in the defense space? Would you agree with that? Is that one of the areas–
SHYAM SANKAR: Well, the current part of the policy is there’s two parts of it. There’s the kind of military doctrine, which is that a human is always accountable for every decision that’s made, and the other part of it is that humans are in- in the decision-making processes that- that relate to lethal force here.
MARGARET BRENNAN: But in terms of the things where you think there should be more guardrails, is the defense space one of them? Biotech, I’ve heard others.
SHYAM SANKAR: All spaces need to- need to have a thought around safety of application.
MARGARET BRENNAN: But that’s still- who does that and how?
SHYAM SANKAR: Well, I think the Department and the Congress should do it with defense. The Department of Transportation should do it with transportation. The FDA should do it with medicines.
MARGARET BRENNAN: But so, watch for that possibly happen. Because you know, I hear from people in the technology space this argument that we need to just go so furiously forward to compete with China. China is pushing for AI to be integrated into all of its industries, but they have the ability to also delay things to get government- governance to catch up, they’ve done that. They’ve tapped the brakes themselves before. So, do you think the U.S. should also consider tapping the brakes in certain areas, just slow things until they have some idea of the guardrails?
SHYAM SANKAR: You know, we have this beautiful, messy process in America, but I would argue, as frustrating as it might look, it’s kind of working. So, we have frontier models now that have not been released because people have realized maybe we’re not ready to release them. That would have been an unimaginable thought a year ago. And so as people are starting to wrap their heads around whether it’s the labs, government- I think it is a consequence of our messy process and the discussions that people are having that we’re evolving how we’re doing this, and the pace is naturally changing.
MARGARET BRENNAN: Let’s talk a little bit about DOD, the Department of War, Department of Defense. They have recently ceased using Anthropic products as part of this dispute over the firm’s insistence on safeguards for military use of AI. I know you’re aware of this case. Anthropic’s Claude was part of the Maven platform, which Palantir uses. How much of an impact will this decision here have, and who replaces them?
SHYAM SANKAR: Well, the- the Department today has a wide range of model providers, everything from XAI and OpenAI to open-weight models like Nemotron from NVIDIA, so I don’t think- I don’t- there’s been no operational impact from that. I think there’s a kind of a broader statement which really has been powering the growth of our commercial business, which is companies are realizing, and it equally applies to the government, that you’re not going to pick any one model provider. That actually, your sovereignty and agency as an institution comes down to your ability to kind of control, decide which models you’re going to use in what context, have the ability to switch models, to be able to control the weights themselves, where you need to feel that control over them. It is your alpha. It is what makes your institution special.
MARGARET BRENNAN: So you don’t know yet who’s going to replace them, but you don’t see a big impact from this kind of decision. It was certainly a huge story for Silicon Valley and the defense space to hear that–
SHYAM SANKAR: It’s already been replaced. It’s already been replaced, and users who are building AI applications get to pick the models that they’re using. It’s a suite of models that are available from, you know, Google, from XAI, from OpenAI. And what was the second part of your question there, sorry?
MARGARET BRENNAN: Yes, I was asking about the impact. It sounds like you don’t think it’s- it’s a big one. Look, I can move on here.
SHYAM SANKAR: No impact.
MARGARET BRENNAN: You have said recently that the war in Iran is the first major conflict where AI played a central role, particularly for targeting. Do you see risks for that?
SHYAM SANKAR: Well, you know, targeting is this- is kind of this very salacious word. Really, targeting is a process, and if you think about military operations, you know, the ability to execute 1000 strikes in a single day would have taken two weeks of planning by 50 people before, and being able to do that with one person in four hours provides you a significant battlefield advantage against your adversaries. So, it’s really not about, you know, people jump immediately to something like AI picking the targets. That’s not what’s happening. It’s really given a set of effects that you want to create- I want to be able to take out the enemy’s air defenses. Okay, what targets relate to that? Okay, how do I then push them through a process where I understand are any of these targets on the no-strike list? Are any of these targets too close to civilian infrastructure? How do you do all the things that are so susceptible to human error under the time crunch- what’s doing in a way that has much higher repeatability and accuracy, so you can get more throughput to the system?
MARGARET BRENNAN: Well, that goes to a broader argument about the elimination of some of the people who were involved in mitigating civilian impact from strikes. That was a Pentagon decision by Secretary Hegseth. In the case- you just talked about the thousands of strikes, and there were so many successful strikes in terms of the targeting in Iran. But there was also, as you know, this horrific one in Minab that killed at least 123 schoolchildren. Bloomberg reported some personnel relied too much on the AI in the Maven system made by Palantir using AI-enabled software. Have you determined what happened there?
SHYAM SANKAR: Well, we should wait for the final report. But what I can say is that reporting is inaccurate.
MARGARET BRENNAN: The reporting from Bloomberg, you’re saying, is inaccurate?
SHYAM SANKAR: Yeah, it’s inaccurate, and actually, the likely cause is human error. And then- and quickly following that, we actually built additional AI agents that did more checking. That’s part of the process we don’t own, but responding to the urgency of the moment, we built additional AI that actually double checks more of the work outside the process that we own to ensure things like that couldn’t happen again.
MARGARET BRENNAN: So you’re suggesting it was outdated intelligence or human error that fed information into the computer system. Is that what you’re saying?
SHYAM SANKAR: Well, we should wait for the final report, but I think we will find–
MARGARET BRENNAN: Do you know when that’ll be released?
SHYAM SANKAR: I don’t.
MARGARET BRENNAN: Because the initial reports have been widely reported on, including by this network. Do you see a harm in that being made public soon, or is it helpful to what you say are improvements you want to make?
SHYAM SANKAR: Well, I- I will defer to the Department on the- on the timing of how- I have no control over that. But yeah, I would love the- the truth to get out there.
MARGARET BRENNAN: But you believe that it was human error. That’s what you’re saying, not what they’re reporting, which is that they were relying too much on the AI.
SHYAM SANKAR: The Bloomberg reporting is wrong. Yeah.
MARGARET BRENNAN: There are 120 House Democrats who wrote to Secretary Hegseth about this particular case and the use of AI in the strike, including Maven. Do you have any intention to go brief Congress about it or to share information?
SHYAM SANKAR: We have given briefings to Congress, and- and they can, you know, Congress is well aware of the capabilities of Maven.
MARGARET BRENNAN: But in terms of this strike and what went wrong, obviously something went wrong.
SHYAM SANKAR: I would defer to the Department on that. That’s- that’s not our place.
MARGARET BRENNAN: But it is because you are such a- Palantir is such- so interwoven with the U.S. government. I mean, it’s billions of dollars of government contracts right now, do you think more information should be briefed to Congress about the benefits along with the risks?
SHYAM SANKAR: Sadly, I’m just a technologist. I mean, on these policy issues, I would defer to others.
MARGARET BRENNAN: Well, I want to quickly ask you before we go about the modernization efforts, which you are part of. I understand you are lieutenant colonel in the Army Reserves in the Army Executive Innovation Corps.
SHYAM SANKAR: Correct.
MARGARET BRENNAN: Which is new.
SHYAM SANKAR: Yeah.
MARGARET BRENNAN: There are also executives from Meta and OpenAI. So, what are you working on there?
SHYAM SANKAR: My principal role is to help with Army human resourcing in the G-1 department here. So, strategies on how do we, you know, the Army is drowning in talent, what we would call green suiters who actually know how to code, and they’re- they’re- they’re exceptional. How do we employ them? You know, they don’t fit in the kind of industrial era model, the Napoleonic codes. Their skills are kind of rank blind. You might have an E-4 who’s an exceptionally good coder. There’s not an obvious career path for them within the Army. The Army doesn’t know how to manage that talent. So, how do we identify who these exceptional people are? How do we protect them? How do we put them on projects that fully utilize their skills and turn them into valuable assets for- for the military?
MARGARET BRENNAN: So that’s your role within this Corps?
SHYAM SANKAR: That’s right.
MARGARET BRENNAN: So for people who look at that and say these big corporations could have a conflict of interest because they’re also doing business with the U.S. government. How do you respond to that?
SHYAM SANKAR: Well, it’s carefully managed by- by the- the lawyers in the Department. We’re only allowed to work on projects where we don’t have conflicts.
MARGARET BRENNAN: And do you think–
SHYAM SANKAR: I would just say, more broadly, I mean it is the greatest honor to be able to serve in our U.S. military. You know, as an immigrant to this country, it’s one of these- My father, my late father, would have loved to have seen me go to a service academy. Life worked out the way it worked out. Arguably, I hope I have more to give back to my country in this capacity at 44 than I did at 24, and it’s a great pride.
MARGARET BRENNAN: Thank you for your time today, Shyam.
SHYAM SANKAR: Thank you.
MARGARET BRENNAN: We’ll be back in a moment.
发表回复