跳到正文
原文
The Verge · AI· Lauren Feiner·· 2 小时前AI 评分45

前美国反垄断司司长 Jonathan Kanter 创办 FluxCap,用 AI 帮原告挑战科技巨头

Jonathan Kanter is using AI to take on Big Tech

文章要点
  • 1前美国司法部反垄断司司长 Jonathan Kanter 创办 FluxCap,售卖帮原告起诉大公司的 AI 法律研究分析工具
  • 2工具基于专有技术与现有大语言模型,可检验论证漏洞,并从海量证据材料中筛查反竞争行为信号
  • 3Kanter 称其相当于 20 多人全天候工作的人力,首批客户为律所 Kressin Powers,正与公私执法方洽谈
  • 4美国反垄断司 2026 财年获拨款 2.45 亿美元、有 400 多名律师,大型州总检察长一个案件可能只有 10 人
决策要点

反垄断诉讼的核心瓶颈是人力与预算,这款工具把大模型用在证据筛查和论证检验上,试图缩小原告与执法方同大企业法务团队的资源差距。

  • 影响:原告与执法方的取证和案卷处理能力可能被放大,大型企业被告面对的证据筛查压力随之上升;对小团队律所和州执法机构,这类工具以软件补上部分人力缺口。
  • 参考:可考虑先盘点自身业务中可能被解读为排他或施压的协议条款与沟通记录;法务团队值得用真实案卷小范围测试此类工具的证据筛查准确率再决定引入;律所可将其作为人手不足案件的补充产能选项。
  • 依据:工具能力与 20 多人人力当量均来自 Kanter 接受 The Verge 采访的口径,暂无独立验证;反垄断司预算与人员数字、Live Nation 案例为报道引述的事实。
正文

When Jonathan Kanter stepped down as head of the Justice Department’s Antitrust Division in 2024, he set out to keep fighting to dismantle illegal monopolies. But this time, he’s continuing the fight in the private sector, and with a perhaps unexpected tool: AI.

Alongside former acting Antitrust Division chief Doha Mekki, Kanter recently launched a new tech startup, FluxCap. The team is stacked with former DOJ staffers and former Palantir engineers, and it’s selling an AI-based legal research and analysis tool that helps plaintiffs build and strengthen cases against the kinds of well-heeled companies Kanter and Mekki took on at the Antitrust Division. FluxCap is launching with a focus on antitrust cases, but it plans to expand into other practice areas as well.

Kanter imagines FluxCap as a “force multiplier” for antitrust watchdogs taking on powerful opponents in Big Tech and elsewhere.

“This, to me, it’s existential”

“This, to me, it’s existential,” he tells The Verge in an interview. “Because it’s the only way that enforcement is going to work in the long term, is if technology and other tools are able to help scarce resources go significantly further than they can today.”

FluxCap’s platform, built off proprietary technology and existing large language models, is primarily intended for research and analysis. It’s trained to pick up the facts and legal framework that will determine whether a case succeeds, then find potential holes in attorneys’ logic. To help attorneys prepare for trial, FluxCap can also help sift through vast piles of discovery materials, searching for signals that a company acted in anticompetitive ways. It’s part of a wave of legal tech tools that are using AI to try to help lawyers do more with less and make the legal process more efficient.

Getting the tool to find relevant evidence was one of the hardest parts, Kanter says. Executives are unlikely to come out and say they’re monopolists, but surrounding context can give a very specific meaning to certain messages. In the antitrust trial against Live Nation-Ticketmaster, for example, the government played a recorded phone call between the Live Nation CEO and a customer who decided to switch to a different ticketing platform. In response, the CEO reminded the customer that a new arena in town could make it harder to deliver as many concerts to his venue. The customer testified he took it as a “not-so-veiled” threat that Live Nation could cut off events programming if he didn’t also accept its ticketing service.

FluxCap’s team of former government lawyers spent months testing the system with existing and hypothetical test cases to see if it could really help build a solid argument. Getting “basic legal tools” (which Kanter calls “BLTs”) to understand that kind of nuance is difficult because “lawyering isn’t about right or wrong answers,” he says. “Both sides can be right and both sides can be wrong. And what you’re essentially trying to do is you’re trying to persuade a court, a jury, a regulator to reach a certain outcome.”

The team counts antitrust boutique firm Kressin Powers as one of its first customers, and Kanter says it’s in talks with both public and private enforcers. “It gives them that virtual capacity so that a small team can take on a large team,” Kanter says, adding that it provides the manpower of roughly “20-plus people working around the clock.”

The seeds of FluxCap began inside the Antitrust Division, where Kanter and his team were already trying to stretch their limited resources with technology. During the Biden administration, the DOJ went to trial against Google twice and brought monopolization suits against Apple and Live Nation, among other actions. But the Division’s resources and staff pale in comparison to companies with multibillion- to trillion-dollar market caps and large legal teams to review documents and mount defenses. The entire Antitrust Division was appropriated $245 million by Congress for fiscal year 2026, with more than 400 attorneys on staff across several issue areas.

“It’s just not realistic to have the resources necessary to take on these big companies”

The imbalance is even greater for state attorneys general, who have pushed forward with antitrust action as the Trump administration has hung back. Earlier this year, state AGs won a case against Live Nation that the DOJ under the Trump administration settled, and states also sued to block the Paramount-Warner Media merger before ultimately settling, when the DOJ declined. “A large state AG might have 10 people working on a case,” Kanter says. “It’s just not realistic to have the resources necessary to take on these big companies.”

During Kanter’s time as antitrust head, he and others tried AI-based BLTs but found them lacking for complex legal research. “Every antitrust lawyer has tried asking questions to ChatGPT or [AI legal research software] Harvey or one of those tools, and they do okay,” Kanter says. “But they miss the nuance.” While Kanter was at the DOJ, the Antitrust Division secured a $45.5 million grant from the government’s Tech Modernization Fund to implement better tools for managing cases and reviewing evidence. Kanter says his team “never had the opportunity to put that to work” before leaving office, but what they’re building now is “even more ambitious and better than I could have imagined when I was at DOJ.”

Tech alone can’t fully even the playing field, Kanter says. “Government is always going to be outmatched, and they desperately need more funding, and desperately need more budget, and more headcount.” Even if that happens, however, private defendants are likely to “far outweigh anything that the government could ever produce, whether it’s federal government or states or private plaintiffs,” he says. “We believe this will be the great equalizer that will allow those teams to go toe to toe with anyone regardless of their size and regardless how deep their pockets are.”

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

  • Lauren Feiner

来源:The Verge · AI · theverge.com