Elon Musk and Pavel Durov, tech billionaires with shared libertarian views and a commitment to free speech on their respective platforms, are at the center of a brewing storm. Durov, head of the messaging app Telegram, recently announced a one-year, $300 million partnership with Musk's artificial intelligence lab, xAI. The deal aims to integrate xAI's Grok chatbot into Telegram, potentially providing Grok with access to Telegram's nearly one billion users. However, the announcement was immediately met with drama when Musk publicly stated on X, "No deal has been signed," casting uncertainty on the collaboration.

This development comes as Elon Musk navigates a period of transition. Having reportedly recently left a post at the White House connected to a "Department of Government Efficiency," the true impact of which remains unclear, Musk is refocusing on his business empire. This empire has increasingly leaned on AI, with his two-year-old venture, xAI, soaring to an $80 billion valuation amidst the generative AI boom. Musk recently consolidated xAI and X (formerly Twitter) under a single entity, xAI Holdings, a move aimed at providing stability to X, which is valued significantly less.

For xAI, the partnership with Telegram, if it materializes, offers a significant opportunity. Access to Telegram's vast user base—double that of X—could translate into substantial revenue through subscriptions for Grok, priced between $3 and $40 per month. Such distribution partnerships have historical precedents for success, such as Dell Inc. pre-installing Microsoft Corp.’s Windows on its PCs, and Apple Inc. earning $20 billion a year from Alphabet Inc. by making Google the default search tool on iPhones.

A critical aspect of the potential deal involves data. Musk might see an opportunity to train Grok further by scraping messages and content from Telegram, leveraging the platform's extensive data to enhance the AI's capabilities, as Grok has already benefited from training on the mass of text on X. However, Pavel Durov has publicly addressed this concern, stating that user privacy is "paramount." He clarified that Grok would only be able to extract messages that users directly share with the bot, seemingly preempting any broader data access attempts by Musk. This public declaration by Durov, made early in the process, might explain his haste in announcing the partnership publicly.

Should the integration proceed this summer, it would mark one of the most profound integrations of an AI tool into a major Western mobile messaging and broadcasting service. Unlike platforms like X, WhatsApp, or Facebook Messenger, which have AI assistants but do not allow AI to edit messages directly, Telegram users could reportedly use Grok to "improve," "expand," or "change tone" of their texts in real-time. A video demo posted by Durov showed Grok transforming a simple typed word "great" into a more elaborate phrase: "Excellent, team effort at its best. Keep up the good work!"

While this functionality might initially appear as another example of AI bringing blandness and vapidity to human conversation, similar to AI-suggested email replies from Apple and Google, its real-time application in mobile text messages pushes technology further into the mechanics behind human voice and intention. More troublingly, this integration has significant implications for a platform like Telegram, notorious for its minimal content rules and for hosting extremist content, conspiracy theories, and being used by groups like neo-Nazi groups in the US to organize rallies, while also being implicated in the promotion of anti-immigration riots in the UK.

Concerns are amplified by the fact that far-right groups have found ways to "jailbreak" AI bots like Grok, getting them to generate content that breaks the rules of other social media firms. The International Centre for Counter-Terrorism singled out Grok as being relatively easy to exploit, citing instances such as generating a photorealistic image of mountains that spell out offensive messages. Musk has yet to publicly respond to criticisms about jailbreaking Grok.

If Grok allows Telegram users to create more persuasive memes and other forms of propaganda at scale, it could make the platform an even more powerful tool for spreading toxicity, from disinformation to hate speech to other odious content. Musk and Durov have kicked off a mini corporate drama of their own, but the real chaos could unfold on the platform itself, and it may not be pretty.