Elon Musk’s xAI Company Accuses OpenAI of Stealing Trade Secrets

In a new lawsuit filed in a California federal court, Elon Musk’s artificial intelligence (AI) startup xAI has accused Sam Altman’s OpenAI of stealing its trade secrets.

xAI makes the generative AI (GenAI) tool called Grok, and OpenAI makes the GenAI tool ChatGPT.

As the complaint explains,

Generative AI refers to systems that can create new content—text, images, music, code, and more—by learning patterns from massive datasets. Unlike traditional AI, which classifies or predicts and gives rote or stilted responses, generative AI produces outputs that are similar to what would be expected from a human in a conversation.

The complaint says that

Grok is a conversational generative AI developed by xAI that is capable of many functions, including (a) performing natural language processing tasks, including answering questions, retrieving information, writing creatively, and assisting with coding, (b) interpreting, editing, and generating images or videos in various styles from fanciful to photorealistic, and also (c) generating natural language audio responses in response to oral prompts from a user.

To protect its confidential and proprietary information, including its trade secrets, the complaint says xAI has implemented a variety of industry standard—and industry leading—practices, such as:

• routinely conducting security awareness training for all employees;
• conducting background checks on employees and contractors who may access xAI data;
• conducting secure development and data handling training for employees with access to sensitive data, after which such employees must complete an assessment to demonstrate understanding;
• employing a team dedicated to information security;
• adopting the NIST 800-171 Rev.3 framework as a baseline for internal security standards;
• achieving SOC 2, Type II compliance;
• securing endpoints, including employee devices, with ongoing patch maintenance and full disk encryption;
• maintaining a formal written information security policy; and
• requiring each employee to enter into an Employee Confidential Information and Invention Assignment Agreement.

These agreements:

• prohibit xAI employees from disclosing, using, or publishing all confidential knowledge, data, or information (“Confidential Information”) belonging to xAI during and after their employment;
• require xAI employees to acknowledge that the employment creates a relationship of confidence and trust with respect to xAI’s Confidential Information, which xAI has a protectable interest therein;
• require xAI employees to return all Confidential Information to xAI and permanently delete any Confidential Information on their personal devices upon termination of their employment;
• require xAI employees to maintain xAI’s confidentiality during and after their employment; and
• require xAI employees to certify their return of Confidential Information upon leaving the company.

According to the complaint,

The desire to win the artificial intelligence (“AI”) race has driven OpenAI to cross the line of fair play. OpenAI violated California and federal law by inducing former xAI employees, including Xuechen Li, Jimmy Fraiture, and a senior finance executive, to steal and share xAI’s trade secrets. By hook or by crook, OpenAI will do anything when threatened by a better innovator, including plundering and misappropriating xAI's technical advancements, source code, and business plans.

Also, according to the complaint,

AI is the defining technology of the modern era. Its development stands to reshape economies and society. Aside from the advent of the internet, the extent of AI’s potential influence on daily life has no parallel in recent history.

As The Guardian reports,

The lawsuit, filed … in California federal court, alleged that OpenAI was engaged in a “deeply troubling pattern” of hiring away former xAI employees to gain access to trade secrets related to its AI chatbot Grok. The company says OpenAI is pursuing unfair advantages in the race to develop AI technology.

The suit claims OpenAI is targeting people with knowledge of xAI’s key technologies and business plans, including xAI’s source code, then inducing them to breach their confidentiality and other obligations to xAI through “unlawful means.”

The complaint says that “This scheme transcends any notions of ordinary employee mobility.”

xAI claims it discovered the alleged scheme while investigating allegations of trade secret theft against former xAI engineer Xuechen Li (an early hire who helped develop xAI’s code), who was accused of taking the company’s confidential information to OpenAI.

According to the complaint,

Li later admitted, in a handwritten confession prepared while represented by counsel, that he had taken, in addition to xAI’s source code, a slide deck specifically pertaining to xAI’s highly sensitive and confidential AI model training and tuning methods.

Also, according to the complaint,

Li uploaded the entire xAI source code base to a personal cloud account. Li then deleted his browser history and system logs to hide evidence of his misdeeds. Indeed, Li later admitted to deleting the system logs to try to cover his tracks.

The next day, OpenAI allegedly offered Li millions of dollars to work for them.

The complaint includes an image of an email sent by xAI’s lawyer to a former xAI executive, accusing them of breaching their confidentiality obligations.

The former employee, whose name is redacted, responded with a crude one-line insult.

According to the Washington Post,

The lawsuit continues a long-standing feud between Musk and the company he co-founded, OpenAI, whose direction he has increasingly criticized after it established an early lead in the AI race. Musk left the company years earlier as it moved in a direction he regarded as for-profit.

As the New York Times reported, xAI also sued Apple and OpenAI for allegedly suppressing the Grok chatbot in the Apple App Store.

According to the Times,

At first, Mr. Musk posted about how well the chatbot was doing in the App Store rankings. But when it didn’t rise to the No. 1 spot, Mr. Musk said on X that Apple had committed “an unequivocal antitrust violation” and added that Apple was “behaving in a manner that makes it impossible for any A.I. company besides OpenAI to reach #1 in the App Store.”

As the Post reports,

Though xAI has played catch-up to competitors in the artificial intelligence race, its footprint has expanded over the past year. The General Services Administration said Thursday that xAI’s Grok models will be available to U.S. agencies over the next 18 months for 42 cents per organization, a rate that draws on a favorite Musk joke.