We caught up with Yannik Schrade - CEO and Co-Founder of Arcium, a parallelized confidential computing network, bringing fast, scaleable and universal encryption. As the lead architect and visionary behind Arcium, Yannik created the project by leveraging his expertise in cryptography with the goal of advancing data security and confidentiality.
A renowned thought leader, Yannik has spoken at several prominent industry conferences, including the 2024 World Economic Forum, where he challenged TradFi leaders on the importance of decentralization, privacy, and trustlessness. He previously founded ShiftScreen, an iOS app, which attracted over 100,000 paying customers, and was a regular top-seller globally. Yannik studied Computer Science and Mathematics at the Technical University of Munich, and he also studied law, adding another dimension to his extensive portfolio.
What are some of the top challenges in data privacy?
Data privacy is a complex topic with various challenges and implications across different verticals. Some key challenges include data security, regulation, user consent and transparency.
What is confidential computing and how does it help with data privacy?
The data lifecycle is typically made up of three parts: data at rest, data in use, or data in transit. So far, cloud providers and other services handling sensitive data have protected data at rest (in storage) or in transit (when it moves across a network). Confidential computing ensures data remains encrypted and protected throughout the data lifecycle, solving the remaining gap to true end-to-end protection.
What are Multiparty computation eXecution Environments (MXEs) and how are they different from Fully Homomorphic Encryption (FHE)?
MXEs are configurable environments for performing secure multi-party computations (MPC). They define the parameters for the confidential computations executed by the Arcium network (a decentralized network of nodes). MXEs also leverage homomorphic encryption, allowing computations to be performed on top of encrypted data. However, MXEs differ from pure FHE as there are limits to the operations that can be performed on top of encrypted data. In addition, MXEs enable the option of decryption, something pure FHE can’t do but which is often desirable in cases where some information is revealed about the computations performed.
Will appropriate use of AI help make blockchain more secure?
It’s rather the other way around. Blockchain can make AI more secure, especially with decentralized confidential computing. In this way, AI models can be trained on encrypted and/or distributed data sets, ensuring the underlying data is never revealed. Decentralized confidential computing can, with federated learning, allow for verifiable model training and prediction pipelines, facilitating the construction of end-to-end encrypted AI. The training of isolated encrypted data results in fully encrypted models combined with confidential prediction and inference within those models. This will lead to more trustless and secure AI as it becomes responsible AI.
At the same time, I think there are models that could be employed on-chain (or their output could be verified on-chain) for secure authentication of users, potentially removing the need for seed phrases. I could imagine biometrics models used with private inference to authenticate users when performing specific on-chain actions.
