Nvidia board probes reports on CEO Jensen Huang

Nvidia's board of directors announced an independent investigation on Sunday (Nov. 8) into reports that company founder and CEO Jensen Huang was the target of a foreign intelligence operation. Separately, the U.S. Defense Advanced Research Projects Agency (DARPA) has opened a program that would use artificial intelligence to find design flaws in electronic markets. And Meta is preparing to bring Muse, its AI assistant, to the United Kingdom, according to the British newspaper The Times. None of the three stories concerns Brazil. What they share is the weight of major technology companies and agencies in the United States and the United Kingdom.

Nvidia: board opens independent investigation

In a Nov. 8 statement, Nvidia's board said the review will be overseen by a committee of independent directors, with the help of independent outside legal counsel. The reason, according to the statement, is recent reporting about a possible foreign intelligence operation aimed at Huang. The statement does not name the outlet that published the reports or the country involved.

The board says U.S. national and economic security is a priority and that Huang fully supports the review and has pledged to cooperate. It also says it has confidence in him and in the leadership team, and promises more information as appropriate. In the board's words, there is no reason to believe national security or company secrets have been compromised, but the review will examine all available information. The statement neither confirms nor details what the reports say. It also gives no timeline and does not spell out the scope of the investigation.

For readers following the industry from Brazil, this is, for now, a governance move at a supplier of the chips used in artificial intelligence. The source says nothing about any effect on products, prices or customers.

DARPA wants AI to find flaws in market design

The program is called Intrinsic Market Resilience (IMR) and was posted on Sept. 25, according to DARPA's own page. The agency describes its goal as developing repair methods that protect critical sectors of the economy from attacks that exploit the public interfaces of modern markets. Abstracts were due Oct. 22, and full proposals are due Dec. 7.

According to a Business Insider report published Nov. 8, the program is looking for artificial intelligence tools that can expose hidden flaws in the rules governing financial and other digital markets before adversaries exploit them. Program manager Matthew Wilding told the outlet that advances in automated reasoning and AI open up new forms of defense, and that IMR focuses on using them to defend the systems. The concern, the report says, is less about break-ins or data theft and more about weaknesses in the design of the markets themselves.

Business Insider reports that DARPA points to two episodes from the past: the California energy market crisis of the early 2000s, which the agency says cost between $40 billion and $45 billion, and the sudden 2010 plunge in U.S. stock markets known as the “flash crash.” Under the plan described in the report, language models would turn regulatory documents and technical specifications into computational models of how markets work. When they could not demonstrate that a market meets a given standard of resilience, the system would generate a concrete example of how it could fail, so that a fix can be proposed. According to the report, the program does not target the kind of manipulation carried out by profit-seekers, but actors whose aim is to cause instability or economic harm.

The focus is on U.S. markets. The sources do not link the program to Brazil.

Meta prepares Muse for the U.K.

The Times reported that Meta is preparing to launch Muse, the company's AI assistant, in the United Kingdom. According to the newspaper, Mark Zuckerberg is aiming for “billions” of users for the agent, which can shop, fill out forms and book appointments online. The full article is behind a paywall. A launch date and price do not appear in the free portion of the text. The sources consulted say nothing about whether Muse will be available in Brazil, and Meta has not confirmed the plan; it is a plan reported by the newspaper.

An assistant that acts on the user's behalf, rather than only answering questions, raises questions about privacy and about responsibility for purchases and sign-ups it makes on its own. The sources do not say how Meta intends to handle those issues.

Read more

Today's summary is in AI Updates at RaisLife. The full edition, with the day's other items, is in the original AI Updates edition. See also the technology page.

Sources

• NVIDIA: statement from the board of directors
• DARPA: Intrinsic Market Resilience (DARPA-PS-26-138)
• Business Insider: DARPA wants AI to find hidden flaws in digital markets
• The Times: Meta prepares U.K. launch of its Muse assistant

Text by Rai, RaisLife's virtual reporter (artificial intelligence), with automated editorial review. Data: technology at RaisLife.