furkanturan

The blog says:

1. I don’t discuss patient cases with an LLM, and

2. I don’t directly use generative models in decision-making.

For the second point I completely agree.

For the first point, I am asking out of curiosity;

- if the AI tools guarantee that whatever you ask will always keep confidential,

- if the AI companies has no way learning them

- if your prompt will be processed under encryption; would your decision change?

Is your concern purely because of confidentiality?

Regarding the SWIFT example, transactions between banks across countries go through SWIFT’s infrastructure and are encrypted for regulatory and security reasons. Banks and financial institutions operating at this scale are also required to use HSMs (Hardware Security Modules) to meet security and compliance standards, ensuring that sensitive data is processed securely.

Suggesting alternative approaches to such large corporations is not straightforward. Their IT and security teams cannot afford to risk handling sensitive financial data in plaintext. No manager would readily approve a solution that offers additional functionality at the cost of deviating, even slightly, from well-established compliance practices. Encryption is therefore fundamental to how these systems operate.

The challenge arises when these institutions need to perform computations or checks on such encrypted data. Decrypting the data, even under controlled conditions and with trained personnel, is operationally complex, introduces additional risk, and can create a significant compliance burden.

This is where FHE offers a compelling solution. We demonstrate how a check can be performed directly on encrypted transactions against a known blacklist, without requiring the underlying transaction data to be decrypted.

P.S. It is great to work with Marc :)

I never thought about whether this was possible, but now that I know it is, I immediately see how useful it could be for my workflow.

That is the key advantage of FHE: you encrypt your data locally, using keys that you own, and never need to share the plaintext—or the keys—with anyone.

But there is more to it. Teams like us at Belfort, we are working to make FHE practical for organizations that want to deploy it for their own use cases, including on infrastructure they control themselves. Even in these setups, FHE can enable applications that would otherwise be blocked by regulatory, privacy, or compliance requirements.

While encrypted AI chatbots seem within reach, we don’t see them as the primary target today. There is a set of untapped applications, such as inter-bank transfers, industrial use cases, and healthcare; where regulations, privacy laws, and compliance requirements prevent institutions from directly accessing or processing sensitive data.

This is where we believe Belfort can make a real difference: enabling organizations to compute on sensitive data while preserving strong privacy guarantees. And FHE doesn’t necessarily mean relying on Google, AWS, or Azure for encrypted compute. Companies can run FHE on their own infrastructure, allowing them to process their clients’ data privately while maintaining control over their systems and keys.

These applications are our initial focus, and over time, we aim to expand the range of use cases that FHE can unlock.

Thanks a lot. We have some road ahead for using encrypted AI for your agentic work. Today, I would not suggest thinking that as the primary target. Instead, there are many untapped applications (e.g. inter-bank transfers, industry, healthcare) where regulations, privacy laws and compliance requirements restrict institutions from touching data. These will initially be our key enablers, and over time we hope to extend the range of applications. At Belfort we are exploring such use cases that will benefit our FHE acceleration.

Yes, it is. But it’s important to recognize that there are different FHE schemes. Zama has invested heavily in TFHE and achieved impressive results. However, time has shown that TFHE is not the best fit for encrypted AI workloads, with CKKS and related schemes gaining an advantage with mathematical and algorithmic advances continue. At Belfort we have foot on both sides; FPGA acceleration for integrating to Zama's TFHE-rs, and GPU acceleration for CKKS.

Count Belfort too. In addition to our GPU acceleration efforts, we have ongoing ASIC initiatives to further accelerate encrypted compute.

Exactly. That is what we have today at Belfort; not enough for making all AI work privacy preserving, but fast enough for many applications, where otherwise unencrypted compute is not acceptable.