Apertus partners with Proton's Lumo AI assistant

Swiss technology company Proton, which specializes in privacy-focused services, and the research team behind the fully open language model Apertus are partnering to advance open and sovereign AI in Switzerland and Europe. Proton is integrating Apertus 1.5 into its AI assistant, Lumo, making it an available option to tens of millions of users and sharing their feedback with the Apertus team to help improve the model.

Lumo is Proton’s privacy-focused AI assistant and an alternative to the leading AI chatbots from the United States and China. Like Apertus, Lumo was developed in Switzerland.

Apertus 1.5 was created by researchers at EPFL and ETH Zurich together with partner institutions. The fully open language model is trained on the Alps supercomputer at CSCS and released under the Apache 2.0 open-source licence, enabling innovation in both research and industry.

Its training data, source code and development process are transparently documented, allowing researchers, public institutions and companies to inspect, adapt and further develop the model. The model was trained on publicly available data, honours website opt-outs, including retroactively, and is designed to remove personal data from its training data at scale.

Valuable feedback from real-world deployment

One of the greatest advantages enjoyed by leading AI companies and language model developers is access to feedback from millions of users, which they use to improve their models. Open models developed in academic settings typically do not benefit from comparable feedback from real-world use.

By integrating Apertus into Lumo, Proton can provide Apertus researchers with feedback data from millions of Lumo users who have given their informed consent. As a result, the Apertus research team will, for the first time, gain access to voluntary user feedback from real-world deployments.

Like all conversations in Lumo, chats with Apertus 1.5 are protected by zero-access encryption and remain private. Users can, however, actively choose to provide anonymised feedback on Apertus’ responses, which can then be shared with the research team to support the model’s further development.

“Real-world feedback is the one ingredient that open models have been missing. Thanks to Proton, Lumo users will be able to voluntarily provide feedback that contributes to Apertus research. For a fully open, publicly developed model, that is a game changer,” says Imanol Schlag, Research Scientist at the ETH AI Center and co-lead of the Apertus project.

Building blocks for a European AI alternative

“Making Apertus available in Lumo gives us valuable insights into how the model performs in real-world settings. Voluntary feedback also helps us identify where it can be made more useful and more robust,” says Martin Jaggi, Professor at EPFL and co-lead of the Apertus project. “It also means that we are enabling a broader community to contribute to the model’s further improvement.”

Proton and the Apertus research team are contributing to greater technological sovereignty in Switzerland and Europe.

“We don’t want to spend the next decade relying on other countries for the AI we use every day,” says Eamonn Maguire, Director of AI at Proton. “Having our own AI isn’t enough. It needs to be competitive so that people actually use it. This partnership with the Apertus team gives Europe a path to get there.”