Proton released Lumo 2.0 — private AI chat with no training on user data

Proton released Lumo 2.0 in July 2026 — the second generation of its own AI chatbot, offered as a private alternative to ChatGPT. The key promise of the product: user dialogues never enter training datasets, and the entire infrastructure complies with European data protection standards.

How Proton Protects Data in Lumo 2.0

Proton has spent more than a decade building a reputation on technically sound privacy: ProtonMail cannot see the contents of emails, ProtonVPN does not keep connection logs. In Lumo 2.0, the same architectural logic is implemented: dialogues with the chatbot are stored encrypted, which excludes access to their contents — including for company employees.

Key service parameters:

• Correspondence is not used for training — neither current nor future versions of the model
• End-to-end encryption of dialogues, similar to ProtonMail
• Server infrastructure hosted in Switzerland — one of the strictest jurisdictions for privacy laws
• Full GDPR compliance — European regulation on personal data protection
• Lumo 2.0 — second generation chatbot by Proton

Who Needs Lumo 2.0 and Why Now?

Most popular AI assistants — ChatGPT, Gemini, Microsoft Copilot — by default can use user dialogues to improve models, unless the user explicitly opts out. Even with training disabled, conversations are stored on servers of American companies, which creates legal uncertainty for organizations operating under European regulation.

The real audience for Lumo 2.0 — journalists, lawyers, doctors, HR specialists and corporate teams under strict industry oversight. For them, a leak of work dialogues through the training dataset means legal liability or reputational damage. A technically sound promise "we physically cannot read your conversations" is in this context more valuable than any performance benchmark.

Proton's approach implements the principle of privacy by design: privacy is built into the product architecture, not a setting that the user must activate themselves. This is precisely the principle underlying GDPR — and it is exactly what Proton consistently implements in every product of its ecosystem.

How Does Lumo Differ from Other Private Chatbots?

There are already several options on the market for those who do not trust major cloud platforms. Local tools — Ollama, LM Studio — run models directly on the user's device: this is maximum control, but requires powerful hardware and technical expertise. Cloud services with declared privacy offer convenience but fall short in brand trust.

Lumo 2.0 occupies a niche between the two poles: familiar cloud UX plus reputation confirmed by years of operation without leaks and without data disclosure at the first demand of authorities. Proton is not an AI laboratory — the company apparently uses a third-party language model with an added layer of private infrastructure. The company does not disclose which LLM exactly powers Lumo under the hood.

What This Means

Lumo 2.0 does not compete with ChatGPT in model capabilities — this is competition in the field of trust. For Proton, launching a chatbot is a logical expansion of the ecosystem: a product with the same value base that made ProtonMail the leading alternative to Gmail for privacy-conscious professionals. As AI tools are increasingly embedded in work processes, demand for "private AI" will only grow.