Your browser does not support the <audio> element.
Pity OpenAI’s lawyers. On February 29th Elon Musk added yet another problem to their list by suing the maker of ChatGPT, a wildly popular artificial-intelligence (AI) chatbot, for breach of contract. The lawsuit argues that the startup was originally set up as a non-profit with the aim of building AI for the “benefit of humanity”, but has since established a profit-making subsidiary and accepted a large investment from Microsoft, a tech titan valued at $3trn, in exchange for exclusive access to its technology. According to Mr Musk, who was an early investor in OpenAI and has since founded a rival, xAI, these actions benefit not humanity but “literally the largest company in the world”.
Whether or not that argument has legal merit is unclear. On March 5th OpenAI published emails in which Mr Musk apparently backs its plan to create a for-profit business. But the case adds to a long list of legal challenges facing the world’s hottest startup. Could these clip its wings?
OpenAI’s problems started in November, when the board ousted Sam Altman, its boss, and Greg Brockman, a co-founder, alleging that Mr Altman had not been “consistently candid in his communications”. Messrs Altman and Brockman were reinstated days later. But the apparent lack of candour raised eyebrows at the Securities and Exchange Commission, America’s markets regulator, which is now investigating the firm to see whether Mr Altman misled investors. In December the New York Times sued OpenAI and Microsoft for copyright infringement, claiming that the startup had used the newspaper’s articles to help train its models. (On March 5th Microsoft filed a motion to dismiss the case, accusing the Times of “doomsday futurology”.) In January America’s Federal Trade Commission launched an antitrust probe into OpenAI’s deal with Microsoft. Regulators in Britain and the EU are considering similar moves.
So far such legal hurdles are not putting off customers and investors. According to SimilarWeb, a data firm, OpenAI clocked 1.6bn visits to its mobile app and website in February, five times what Gemini, Google’s rival AI, has managed. Surveys of large companies invariably find OpenAI and Microsoft to be the most popular purveyors of AI tools. In December Openai was making an annualised revenue of $2bn, up from virtually nothing the year before. Venture capitalists report that the startups they back see its models as best on price and performance. This week it emerged that Temasek, a Singaporean sovereign-wealth fund, wants a stake in the firm.
Even so, the mounting legal action may hobble OpenAI in the long run. Brendan Burke of PitchBook, a research firm, argues that OpenAI’s lawyers may be worried about the additional legal risks that new models could create. That in turn could lead OpenAI once again to push back the release of GPT-5, expected to be the world’s most powerful AI yet. The delay would give rivals an advantage.
For one thing, it gives competitors time to catch up technologically. In mid-February Google updated its Gemini model. The search giant claims that the new version is able to process far more data than GPT-4 (though it has also been criticised for being too “woke”). Two weeks later Mistral, a French challenger, launched a small but high-performing model. On March 4th Anthropic, an AI startup backed by Google and Amazon, released its latest AI, Claude 3, claiming that it outperforms GPT-4 on a number of tasks (including the Multistate Bar Examination for lawyers).
This in turn is leading businesses to think more carefully about diversifying their AI toolboxes. Startups and corporations are already designing software to make it easy to switch between OpenAI’s current top model, GPT-4, and alternatives. That is partly a way to minimise costs by using cheaper, less powerful models for simpler tasks, but also insurance in case one model-maker runs into trouble. Paul Daugherty of Accenture, a consultancy, says one of the most common questions company bosses ask about AI is which model they should use. This suggests that OpenAI is no longer the default choice. Even Microsoft is spreading its bets. Last month it announced a partnership with Mistral, whose open-source models will be available on Microsoft’s Azure cloud.
Court cases and regulatory probes are, then, an unwelcome distraction for OpenAI. As for Mr Altman, he seems confident he can multitask. So confident, in fact, that he is reportedly in talks with the government of the United Arab Emirates and SoftBank, a Japanese tech-investing powerhouse, among others, to build a $7trn AI-chip manufacturer. ■
To stay on top of the biggest stories in business and technology, sign up to the Bottom Line, our weekly subscriber-only newsletter.
