`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`IRELL & MANELLA LLP
`Morgan Chu (SBN 70446)
`Alan Heinrich (SBN 212782)
`Iian Jablon (SBN 205458)
`Abigail Sellers (SBN 342380)
`Justin Koo (SBN 351547)
`Henry White (SBN 351549)
`1800 Avenue of the Stars, Suite 900
`Los Angeles, CA 90067
`Telephone: (310) 277-1010
`Fax: (310) 203-7199
`mchu@irell.com
`aheinrich@irell.com
`ijablon@irell.com
`asellers@irell.com
`jkoo@irell.com
`hwhite@irell.com
`
`Attorneys for Plaintiff Elon Musk
`
`ELECTRONICALLY
`F I L E D
`
`Superior Court of California,
`County of San Francisco
`02/29/2024
`Clerk of the Court
`BY: KEVIN DOUGHERTY
`Deputy Clerk
`
`
`
`SUPERIOR COURT OF CALIFORNIA
`
`CGC-24-612746
`IN AND FOR THE COUNTY OF SAN FRANCISCO
`
`ELON MUSK, an individual,
`Plaintiff,
`
`Case No.:
`
`vs.
`SAMUEL ALTMAN, an individual, GREGORY
`BROCKMAN, an individual, OPENAI, INC., a
`corporation, OPENAI, L.P., a limited
`partnership, OPENAI, L.L.C., a limited liability
`company, OPENAI GP, L.L.C., a limited
`liability company, OPENAI OPCO, LLC, a
`limited liability company, OPENAI GLOBAL,
`LLC, a limited liability company, OAI
`CORPORATION, LLC, a limited liability
`company, OPENAI HOLDINGS, LLC, a limited
`liability company, and DOES 1 through 100,
`inclusive,
`
`Defendants.
`
`[UNLIMITED JURISDICTION]
`
`COMPLAINT FOR (1) BREACH OF
`CONTRACT, (2) PROMISSORY
`ESTOPPEL, (3) BREACH OF FIDUCIARY
`DUTY, (4) UNFAIR COMPETITION
`UNDER CAL. BUS. & PROF. CODE
`§§ 17200 ET SEQ., AND (5) ACCOUNTING
`
`
`DEMAND FOR JURY TRIAL
`
`
`
`
`
`
`COMPLAINT
`
`
`
`
`
`
`belief:
`
`Plaintiff, ELON MUSK (hereafter “Plaintiff”) alleges the following upon information and
`
`PARTIES
`1.
`Prior to 2019, Plaintiff was an individual residing in California. Plaintiff is a resident
`of Texas since 2019.
`2.
`On information and belief, Plaintiff alleges that Samuel Altman is a resident of the
`County of San Francisco, State of California.
`3.
`On information and belief, Plaintiff alleges that Gregory Brockman is a resident of
`the County of San Francisco, State of California.
`4.
`OpenAI, Inc. is a registered non-profit organization incorporated under the laws of
`Delaware on December 8, 2015. OpenAI, Inc. is registered as a foreign corporation with the
`California Secretary of State and has its principal place of business at 3180 18th Street, San
`Francisco, CA 94110.
`5.
`OpenAI, L.P. is a limited partnership formed under the laws of Delaware on
`September 19, 2018, originally as SummerSafe, L.P. OpenAI, L.P. is registered as a foreign limited
`partnership with the California Secretary of State and has its principal place of business at 3180 18th
`Street, San Francisco, CA 94110.
`6.
`OpenAI, L.L.C. is a limited liability company formed in Delaware on September 17,
`2020. OpenAI, L.L.C. maintains its principal place of business in California.
`7.
`OpenAI GP, L.L.C. is a limited liability company formed in Delaware on September
`19, 2018. OpenAI GP, L.L.C is registered as a foreign limited liability company registered with the
`California Secretary of State and has its principal place of business at 3180 18th Street, San
`Francisco, CA 94110.
`8.
`OpenAI OpCo, LLC is a limited liability company formed in Delaware on September
`19, 2018. OpenAI OpCo, LLC is registered as an out-of-state limited liability company with the
`California Secretary of State and has it principal place of business at 1960 Bryant Street, San
`Francisco, CA 94110.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`
`– 2 –
`COMPLAINT
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`9.
`OpenAI Global, LLC is a limited liability company formed in Delaware on
`December 28, 2022. OpenAI Global, LLC is registered as an out-of-state limited liability company
`with the California Secretary of State and has it principal place of business at 1960 Bryant Street,
`San Francisco, CA 94110.
`10.
`OAI Corporation, LLC is a limited liability company formed in Delaware. OAI
`Corporation, LLC maintains its principal place of business in California.
`11.
`OpenAI Holdings, LLC is a limited liability company formed in Delaware on March
`17, 2023. OpenAI Holdings, LLC is registered as an out-of-state limited liability company with the
`California Secretary of State and has it principal place of business at 1960 Bryant Street, San
`Francisco, CA 94110.
`12.
`Hereinafter, “OpenAI, Inc.” is used solely to refer to the non-profit entity or non-
`profit arm, while “OpenAI” is used generally to refer to OpenAI, Inc., OpenAI, L.P., OpenAI,
`L.L.C., OpenAI GP, L.L.C., OpenAI OpCo, LLC, OpenAI Global, LLC, OAI Corporation, LLC,
`and/or OpenAI Holdings, LLC.
`13.
`Plaintiff is currently unaware as to the names and identities of Doe 1 through Doe
`
`100.
`
`JURISDICTION AND VENUE
`14.
`On information and belief, Plaintiff alleges that many of the occurrences,
`representations, and events upon which this action is based took place in County of San Francisco,
`State of California, where the vast majority of Defendants reside or have their principal place of
`business, and the vast majority of the occurrences, representations, and events upon which this
`action is based took place in the State of California.
`GENERAL ALLEGATIONS
`Overview Of The Case
`A.
`The Risk Of Artificial General Intelligence
`15.
`Over the course of the 20th century, the United States gradually shifted from a
`primarily human labor-based economy to a primarily human knowledge-based economy, with
`economic value increasingly created primarily by human intelligence. As the century progressed,
`
`
`
`
`– 3 –
`COMPLAINT
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`another paradigm shift was already underway: value creation through artificial intelligence (AI).
`Early AI programs were capable of outperforming humans in certain discrete tasks. Almost
`immediately after the invention of the programmable computer, AI could show super-human
`performance on a highly formalized problem like finding the fastest path through a network of roads.
`It took longer for AI to reach superiority for problems requiring more creativity. In 1996, IBM’s
`Deep Blue AI program beat Gary Kasparov, the then-world champion in chess. These programs,
`while useful, were essentially one-trick-ponies—their intelligence was not general. Deep Blue had
`massive arrays of processors that could only play chess. Algorithms for path finding could solve a
`maze or route a car, but they could not paint a painting.
`16.
`Starting in the late 2000s and early 2010s, an older algorithm called “deep learning”
`became practical to implement on low-cost hardware for the first time. This caused an almost
`overnight revolution in performance across nearly all AI projects. New, top of class algorithms were
`developed for turning speech into text, translating between languages, and recognizing what kind of
`food is shown in a photo. One of the hallmarks of deep learning is that algorithms do not need to be
`designed with significant knowledge of the task at hand. They learn each task from training
`examples, essentially programming themselves. This means that they are far more general-purpose
`than earlier systems like Deep Blue.
`17.
`As deep learning algorithms became increasingly sophisticated, some of the world’s
`leading AI researchers set their sights on what has come to be called Artificial General Intelligence
`(AGI). The basic concept of AGI is a general purpose artificial intelligence system—a machine
`having intelligence for a wide variety of tasks like a human.
`18. Mr. Musk has long recognized that AGI poses a grave threat to humanity—perhaps
`the greatest existential threat we face today. His concerns mirrored those raised before him by
`luminaries like Stephen Hawking and Sun Microsystems founder Bill Joy. Our entire economy is
`based around the fact that humans work together and come up with the best solutions to a hard task.
`If a machine can solve nearly any task better than we can, that machine becomes more economically
`useful than we are. As Mr. Joy warned, with strong AGI, “the future doesn’t need us.” Mr. Musk
`
`
`
`
`– 4 –
`COMPLAINT
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`publicly called for a variety of measures to address the dangers of AGI, from voluntary moratoria
`to regulation, but his calls largely fell on deaf ears.
`19.
`But where some like Mr. Musk see an existential threat in AGI, others see AGI as a
`source of profit and power.
`20.
`In 2014, Google acquired DeepMind, a research group focused on deep learning.
`One of DeepMind’s initial developments was AlphaZero, a chess playing algorithm. Unlike
`previous algorithms, however, AlphaZero used “reinforcement learning,” wherein the program
`learns to play chess by playing itself with different versions of the software. It starts by playing
`randomly, with no understanding of the game’s strategy. When one version of the software wins a
`game against another, the winning program’s internal pathways are “reinforced” and the process
`repeats.
`21.
`AlphaZero rapidly became the strongest chess playing system in the world.
`Shockingly, it was also announced that the same program was also the strongest in the world for
`playing two other extremely difficult games. In Google/DeepMind’s words, “Starting from random
`play, and given no domain knowledge except the game rules, AlphaZero achieved within 24 hours
`a superhuman level of play in the games of chess and shogi (Japanese chess) as well as Go, and
`convincingly defeated a world-champion program in each case.”
`22. With the DeepMind team, Google immediately catapulted to the front of the race for
`AGI. Mr. Musk was deeply troubled by this development. He believed (and still does) that in the
`hands of a closed, for-profit company like Google, AGI poses a particularly acute and noxious
`danger to humanity. In 2014, it was already difficult enough to compete with Google in its core
`businesses. Google had collected a uniquely large set of data from our searches, our emails, and
`nearly every book in our libraries. Nevertheless, up to this point, everyone had the potential to
`compete with Google through superior human intelligence and hard work. AGI would make
`competition nearly impossible.
`B.
`The Founding Agreement Of OpenAI, Inc.
`23. Mr. Altman purported to share Mr. Musk’s concerns over the threat posed by AGI.
`In 2015, Mr. Altman wrote that the “[d]evelopment of superhuman machine intelligence (SMI) is
`
`
`
`
`– 5 –
`COMPLAINT
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`probably the greatest threat to the continued existence of humanity. There are other threats that I
`think are more certain to happen . . . but are unlikely to destroy every human in the universe in the
`way that SMI could.” Later that same year, Mr. Altman approached Mr. Musk with a proposal: that
`they join forces to form a non-profit AI lab that would try to catch up to Google in the race for AGI,
`but it would be the opposite of Google.
`24.
`Together with Mr. Brockman, the three agreed that this new lab: (a) would be a non-
`profit developing AGI for the benefit of humanity, not for a for-profit company seeking to maximize
`shareholder profits; and (b) would be open-source, balancing only countervailing safety
`considerations, and would not keep its technology closed and secret for proprietary commercial
`reasons (The “Founding Agreement”). Reflecting the Founding Agreement, Mr. Musk named this
`new AI lab “OpenAI,” which would compete with, and serve as a vital counterbalance to,
`Google/DeepMind in the race for AGI, but would do so to benefit humanity, not the shareholders
`of a private, for-profit company (much less one of the largest technology companies in the world).
`25.
`The Founding Agreement was also memorialized, among other places, in OpenAI,
`Inc.’s December 8, 2015 Certificate of Incorporation, which affirmed that its “resulting technology
`will benefit the public and the corporation will seek to open source technology for the public benefit
`when applicable. The corporation is not organized for the private gain of any person.” Ex. 1 at 1.
`The Certificate of Incorporation further affirmed that all of the corporation’s property was
`“irrevocably dedicated” to these agreed purposes. Id.
`26.
`In reliance on the Founding Agreement, which Mr. Altman, Mr. Brockman, and
`OpenAI, Inc. reaffirmed with Mr. Musk on multiple occasions, Mr. Musk was a moving force
`behind the creation of OpenAI, Inc. contributing a majority of its funding in its first several years,
`advising on research directions, and most importantly, recruiting some of the world’s leading
`scientists and engineers to work at the non-profit venture, including Chief Scientist Ilya Sutskever.
`Recruiting for OpenAI, Inc. was a Herculean task in the face of relentless recruiting efforts by
`Google/DeepMind, coupled with the lavish compensation Google/DeepMind offered. There would
`have been no OpenAI, Inc. without Mr. Musk’s founding contributions and early leadership. Mr.
`
`
`
`
`– 6 –
`COMPLAINT
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`Musk continued to make contributions to OpenAI, Inc. from its founding through September 14,
`2020.
`
`27.
`OpenAI’s initial research was performed in the open, providing free and public
`access to designs, models, and code. When OpenAI, Inc. researchers discovered that an algorithm
`called “Transformers,” initially invented by Google, could perform many natural language tasks
`without any explicit training, entire communities sprung up to enhance and extend the models
`released by OpenAI, Inc. These communities spread to open-source, grass-roots efforts and
`commercial entities alike.
`28. Mr. Altman became OpenAI, Inc.’s CEO in 2019. On September 22, 2020, OpenAI
`entered into an agreement with Microsoft, exclusively licensing to Microsoft its Generative Pre-
`Trained Transformer (GPT)-3 language model. However, OpenAI published a detailed paper
`describing the internals and training data for GPT-3, enabling the community to create similar
`models themselves. And, most critically, the Microsoft license only applied to OpenAI’s pre-AGI
`technology. Microsoft obtained no rights to AGI. And it was up to OpenAI, Inc.’s non-profit Board,
`not Microsoft, to determine when OpenAI attained AGI.
`C.
`The 2023 Breach Of The Founding Agreement
`29.
`In 2023, Defendants Mr. Altman, Mr. Brockman, and OpenAI set the Founding
`Agreement aflame.
`30.
`In March 2023, OpenAI released its most powerful language model yet, GPT-4.
`GPT-4 is not just capable of reasoning. It is better at reasoning than average humans. It scored in
`the 90th percentile on the Uniform Bar Exam for lawyers. It scored in the 99th percentile on the
`GRE Verbal Assessment. It even scored a 77% on the Advanced Sommelier examination. At this
`time, Mr. Altman caused OpenAI to radically depart from its original mission and historical practice
`of making its technology and knowledge available to the public. GPT-4’s internal design was kept
`and remains a complete secret except to OpenAI—and, on information and belief, Microsoft. There
`are no scientific publications describing the design of GPT-4. Instead, there are just press releases
`bragging about performance. On information and belief, this secrecy is primarily driven by
`commercial considerations, not safety. Although developed by OpenAI using contributions from
`
`
`
`
`– 7 –
`COMPLAINT
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`Plaintiff and others that were intended to benefit the public, GPT-4 is now a de facto Microsoft
`proprietary algorithm, which it has integrated into its Office software suite.
`31.
`Furthermore, on information and belief, GPT-4 is an AGI algorithm, and hence
`expressly outside the scope of Microsoft’s September 2020 exclusive license with OpenAI. In this
`regard, Microsoft’s own researchers have publicly stated that, “[g]iven the breadth and depth of
`GPT-4’s capabilities, we believe that it could reasonably be viewed as an early (yet still incomplete)
`version of an artificial general intelligence (AGI) system.” Moreover, on information and belief,
`OpenAI is currently developing a model known as Q* (Q star) that has an even stronger claim to
`AGI. As noted, Microsoft only has rights to certain of OpenAI’s pre-AGI technology. But for
`purposes of the Microsoft license, it is up to OpenAI, Inc.’s Board to determine whether OpenAI
`has attained AGI, and a Board coup took place in November 2023. On November 17, 2023, OpenAI,
`Inc.’s Board fired Mr. Altman after losing “confidence in his ability to continue leading OpenAI”
`because “he was not consistently candid with the board.” In a series of stunning developments
`spanning the next several days, Mr. Altman and Mr. Brockman, in concert with Microsoft, exploited
`Microsoft’s significant leverage over OpenAI, Inc. and forced the resignation of a majority of
`OpenAI, Inc.’s Board members, including Chief Scientist Ilya Sutskever. Mr. Altman was reinstated
`as CEO of OpenAI, Inc. on November 21. On information and belief, the new Board members were
`hand-picked by Mr. Altman and blessed by Microsoft. The new Board members lack substantial AI
`expertise and, on information and belief, are ill equipped by design to make an independent
`determination of whether and when OpenAI has attained AGI—and hence when it has developed
`an algorithm that is outside the scope of Microsoft’s license.
`32.
`These events of 2023 constitute flagrant breaches of the Founding Agreement, which
`Defendants have essentially turned on its head. To this day, OpenAI, Inc.’s website continues to
`profess that its charter is to ensure that AGI “benefits all of humanity.” In reality, however, OpenAI,
`Inc. has been transformed into a closed-source de facto subsidiary of the largest technology company
`in the world: Microsoft.1 Under its new Board, it is not just developing but is actually refining an
`AGI to maximize profits for Microsoft, rather than for the benefit of humanity. Its technology,
`
`1 Microsoft’s market cap as of February 28, 2024 was $3.03 trillion.
`
`
`
`
`– 8 –
`COMPLAINT
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`including GPT-4, is closed-source primarily to serve the proprietary commercial interests of
`Microsoft. Indeed, as the November 2023 drama was unfolding, Microsoft’s CEO boasted that it
`would not matter “[i]f OpenAI disappeared tomorrow.” He explained that “[w]e have all the IP
`rights and all the capability.” “We have the people, we have the compute, we have the data, we have
`everything.” “We are below them, above them, around them.”
`33.
`This case is filed to compel OpenAI to adhere to the Founding Agreement and return
`to its mission to develop AGI for the benefit of humanity, not to personally benefit the individual
`Defendants and the largest technology company in the world.
`Detailed Allegations
`A. Mr. Musk’s Concerns Over AGI Falling Into The Wrong Hands
`34.
`In 2012, Elon Musk met Demis Hassabis, the co-founder of DeepMind, a for-profit
`artificial intelligence company. On or about this time, Mr. Musk and Mr. Hassabis met at SpaceX’s
`factory in Hawthorne, California where Mr. Musk and Mr. Hassabis discussed the greatest threats
`facing society. During this conversation, Mr. Hassabis emphasized the potential dangers that the
`advancement of AI presents to society.
`35.
`Following this conversation with Mr. Hassabis, Mr. Musk became increasingly
`concerned about the potential of AI to become super-intelligent, surpass human intelligence, and
`threaten humanity. Indeed, Mr. Musk was not the only person afraid about the dangers of AI and
`the research being conducted at DeepMind. It has been reported that following a meeting with Mr.
`Hassabis and investors in DeepMind, one of the investors remarked that the best thing he could have
`done for the human race was shoot Mr. Hassabis then and there.
`36. Mr. Musk began discussing AI and DeepMind with those in his orbit, such as Larry
`Page, then-CEO of Google’s parent company Alphabet, Inc. Mr. Musk would frequently raise the
`dangers of AI in his conversations with Mr. Page, but to Mr. Musk’s shock, Mr. Page was
`unconcerned. For example, in 2013, Mr. Musk had a passionate exchange with Mr. Page about the
`dangers of AI. He warned that unless safeguards were put in place, “artificial intelligence-systems
`might replace humans, making our species irrelevant or even extinct.” Mr. Page responded that
`would merely “be the next stage of evolution,” and claimed Mr. Musk was being a “specist”—that
`
`
`
`
`– 9 –
`COMPLAINT
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`he favored the human species over intelligent machines. Mr. Musk responded, “Well, yes, I am pro-
`human.”
`37.
`At the end of 2013, Mr. Musk learned to his grave concern that Google was planning
`to acquire DeepMind. At the time, DeepMind was one of the most advanced AI companies in the
`industry. Thus, Mr. Musk was deeply concerned that DeepMind’s AI technology would be in the
`hands of someone who viewed it and its power so cavalierly, and could hide its design and
`capabilities behind closed doors.
`38.
`In an effort to prevent this powerful technology from falling into Google’s hands,
`Mr. Musk and Luke Nosek, a co-founder of PayPal, attempted to put together funding to buy
`DeepMind. This effort culminated in an hour-long call wherein Mr. Musk and Mr. Nosek made one
`last effort to convince Mr. Hassabis not to sell DeepMind to Google. Mr. Musk told Mr. Hassabis
`that “[t]he future of AI should not be controlled by Larry [Page].”
`39. Mr. Musk and Mr. Nosek’s effort was unsuccessful. It was reported in January 2014
`that DeepMind would be acquired by Google. However, this did not deter Mr. Musk from continuing
`to ensure that AI was developed and practiced safely.
`40.
`Following Google’s acquisition of DeepMind, Mr. Musk began “hosting his own
`series of dinner discussions on ways to counter Google and promote AI safety.” Mr. Musk also
`reached out to President Barack Obama to discuss AI and AI safety. In 2015, Mr. Musk and
`President Obama had a meeting during which Mr. Musk explained the dangers of AI and advocated
`for regulation. Mr. Musk felt that President Obama understood the dangers of AI, but regulation
`never came.
`41.
`Despite these setbacks, Mr. Musk continued to advocate for safe AI practices. In
`2015, it appeared that Mr. Musk may have found someone who understood his concerns about AI
`and his desire to keep the first AGI out of the hands of a private company like Google: Defendant
`Sam Altman.
`42.
`At this time, Mr. Altman was the president of Y Combinator, a start-up accelerator
`in Silicon Valley. Before that, Mr. Altman had been involved in various startup ventures.
`
`
`
`
`– 10 –
`COMPLAINT
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`43. Mr. Altman appeared to share Mr. Musk’s concerns surrounding AI. In public blog
`posts dating back to 2014, Mr. Altman stated that AGI, if made, would “be the biggest development
`in technology ever.” Mr. Altman pointed out that there are many companies making strides towards
`achieving AGI, but acknowledged the unfortunate reality that the “good ones are very secretive
`about it.”
`44.
`On February 25, 2015, Mr. Altman also expressed his concern surrounding the
`development of what he referred to as “superhuman machine intelligence” which he identified as
`“probably the greatest threat to the continued existence of humanity” and emphasized that “as a
`human programmed to survive and reproduce, I feel we should fight it.” Further, Mr. Altman
`criticized those who believed that “superhuman machine intelligence” was dangerous but dismissed
`it as “never going to happen or definitely very far off.” He accused them of engaging in “sloppy,
`dangerous thinking.”
`45.
`Indeed, in early March 2015, Mr. Altman extolled the importance of government
`regulation as a means to ensure AI is created safely and suggested that “a group of very smart people
`with a lot of resources” likely involving “US companies in some way” would be the most probable
`group to achieve “superhuman machine intelligence” first.
`46.
`Later that month, Mr. Altman reached out to Mr. Musk to inquire whether he would
`be interested in drafting an open letter to the United States Government regarding AI. The two began
`preparing a letter and approaching those of influence in the technology and AI sectors about signing.
`It did not take long before those across the industry heard rumors of the letter.
`47.
`For example, in April of 2015, Mr. Hassabis reached out to Mr. Musk stating that he
`had heard from multiple sources that Mr. Musk was drafting a letter addressed to the President
`calling for regulation of AI. Mr. Musk defended the idea of the regulation of AI to Mr. Hassabis,
`stating: “If done well, it may very well accelerate AI in the long term. Without the public comfort
`that regulatory oversight provides, there could very well be a situation where an AI causes great
`harm and thereafter AI research is banned as dangerous to public safety.”
`48.
`Five days after Mr. Hassabis reached out to Mr. Musk about the open letter regarding
`AI, Mr. Hassabis announced the first meeting of the Google DeepMind AI Ethics Board, a board
`
`
`
`
`– 11 –
`COMPLAINT
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`that Google and DeepMind promised to establish two years earlier when Google acquired
`DeepMind. Mr. Musk was asked to be a member of the board and offered for the first meeting to be
`hosted at SpaceX in Hawthorne, California. It became clear to Mr. Musk after the first meeting that
`this board was not a serious endeavor, but was rather a façade to try to slow down any regulation
`of AI.
`
`49.
`The Open Letter was later published on October 28, 2015, and signed by over eleven
`thousand individuals, including Mr. Musk, Stephen Hawking, and Steve Wozniak.
`B.
`OpenAI, Inc.’s Founding Agreement
`50.
`On May 25, 2015, Mr. Altman emailed Mr. Musk, writing that he had “[b]een
`thinking a lot about whether it’s possible to stop humanity from developing AI. I think the answer
`is almost definitely not. If it’s going to happen, it seems like it would be good for someone other
`than Google to do it first.” Mr. Altman had an idea: that Y Combinator start a “Manhattan Project”
`for AI. (Unfortunately, it turns out that the “Manhattan Project” is a moniker that may be all too
`apt.) He proposed that “we could structure it so that the tech belongs to the world via some sort of
`nonprofit but the people working on it get startup-like compensation if it works. Obviously we’d
`comply with/aggressively support all regulation.” Mr. Musk responded, “Probably worth a
`conversation.”
`51.
`After further communications, Mr. Altman emailed Mr. Musk on June 24, 2015 with
`a detailed proposal for this new “AI lab.” “The mission would be to create the first general AI and
`use it for individual empowerment—ie, the distributed version of the future that seems the safest.
`More generally, safety should be a first-class requirement.” “The technology would be owned by
`the foundation and used ‘for the good of the world’[.]” He proposed that they start with a group of
`7-10 people and expand from there. He also proposed a governance structure. Mr. Musk responded,
`“Agree on all.” Ex. 2 at 1.
`52.
`Soon thereafter, Mr. Altman began recruiting others to help with the development of
`the project. Notably, Mr. Altman approached Gregory Brockman to help with the project.
`53.
`In November 2015, Mr. Altman put Mr. Brockman in communication with Mr. Musk
`via email. Regarding the project, Mr. Brockman told Mr. Musk, “I hope for us to enter the field as
`
`
`
`
`– 12 –
`COMPLAINT
`
`
`
`a neutral group looking to collaborate widely and shift the dialog towards being about humanity
`winning rather than any particular group or company. (I think that’s the best way to bootstrap
`ourselves into being a leading research institution.).” Optimistic about the possibility of a neutral AI
`research group focused on humanity rather than profit for any particular individual or group,
`Mr. Musk told Mr. Brockman that he would commit funding.
`54. Mr. Musk came up with the name of the new lab, a name reflecting the Founding
`Agreement: “Open AI Institute,” or simply, “OpenAI.”
`55. With these principles of the Founding Agreement in mind, Mr. Musk joined forces
`with Mr. Altman and Mr. Brockman to formalize and launch the project. Mr. Musk was actively
`involved in the project, even prior to it being publicly announced. For example, Mr. Musk advised
`Mr. Brockman on compensation packages for employees, sharing with Mr. Brockman his strategies
`for compensation and retaining talent.
`56.
`On December 8, 2015, a Certificate of Incorporation for OpenAI, Inc. was filed with
`the Delaware Secretary of State. The Certificate memorialized in writing the Founding Agreement:
`THIRD: This Corporation shall be a nonprofit corporation organized exclusively for
`charitable and/or educational purposes within the meaning of section 501(c)(3) of
`the Internal Revenue Code of 1986, as amended, or the corresponding provision of
`any future United States Internal Revenue law. The specific purpose of this
`corporation is to provide funding for research, development and distribution of
`technology related to artificial intelligence. The resulting technology will benefit the
`public and the corporation will seek to open source technology for the public benefit
`when applicable. The corporation is not organized for the private gain of any
`person. . . .
`
`FIFTH: The property of this corporation is irrevocably dedicated to the purposes in
`Article THREE hereof and no part of the net income or assets of this corporation
`shall ever inure to the benefit of any director, officer or member thereof or to the
`benefit of any private person. Upon the dissolution or winding up of this corporation,
`its assets remaining after payment, or provision for payment, of all debts and
`liabilities of this corporation shall be distributed to a nonprofit fund, foundation, or
`corporation which is organized and operated exclusively for charitable, educational
`and/or religious purposes and which has established its tax exempt status under
`Internal Revenue Code section 501(c)(3), or the corresponding section of any future
`federal tax code, or shall be distributed to the federal government, or to a state or
`local government, for a public purpose.
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`
`– 13 –
`COMPLAINT
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`57.
`OpenAI, Inc. was publicly announced on December 11, 2015. In the announcement,
`Mr. Musk and Mr. Altman were named as co-chairs and Mr. Brockman was named as the CTO. The
`announcement emphasized that OpenAI was designed to “benefit humanity,” and its research would
`be “free from financial obligation”:
`OpenAI is a non-profit artificial intelligence research company. Our goal is to
`advance digital intelligence in the way that is most likely to benefit humanity as a
`whole, unconstrained by a need to generate financial return. Since our research is
`free from financial obligations, we can better focus on a positive human impact.
`
`Ex. 3 at 1.
`C. Mr. Musk’s Crucial Role In Getting OpenAI, Inc. Off The Ground
`58.
`In an email on the day of the public announcement, Mr. Musk wrote that “[o]ur most
`important consideration is recruitment of the best people.” He pledged that helping with the
`recruiting effort would be his “absolute top priority 24/7.” He acknowledged that “[w]e are
`outmanned and outgunned by a ridiculous margin by organizations you know well, but we have
`right on our side and that counts for a lot. I like the odds.” Mr. Musk brought to bear his connections,