The promise of the data economy is that by unlocking the insight hidden in huge volumes of information, we can take giant steps forward – commercially, if you run a business, but also for society. However, there is a problem. Many people feel very uncomfortable with their data being made available to multiple unknown parties; indeed, in many industries and countries, the law explicitly forbids this. How, then, do those trying to secure insight get their hands on the precious data required?
“The answer is sharing without sharing,” says Juan Troncoso, the co-founder and CEO of Swiss start-up Tune Insight, which is today announcing the successful completion of a $3.4 million funding round. “We need technologies that enable organisations to interrogate large data sets held elsewhere that don’t compromise the privacy or security of that data.”
This is the challenge that Tune Insight set out to solve when the business was launched in 2021. The company uses a technique known as “multi-party homomorphic encryption” to move past privacy and security issues. Effectively, the technology makes it possible to ask questions of another organisation’s data – or multiple organisations’ data – without having to decrypt that data. The interrogator has access to the data in order to search for insight, but there are no privacy or security issues because the data remains encrypted.
The idea was originally developed at the Swiss Federal Institute of Technology following an approach from a number of Swiss hospitals. They recognised that sharing their patient data – and accessing more data from other hospitals – was crucial in the age of personalised medicine, where physicians aim to provide far more bespoke treatment plans. But they did not want – and were not allowed – to do anything to compromise patient privacy.
In a healthcare setting, the impact of Tune Insight’s solution to this problem could have huge societal benefits. An oncologist, say, who is searching for the right way to treat a cancer patient according to their specific symptoms can interrogate the database to find out what has worked best on other patients with exactly the same symptoms. That insight is available without any identification of individual patients – the system simply returns a result.
“These organisations are desperate to collaborate, but they need more private ways to do it,” says Troncoso. “And the more we can build out this model, the more sophisticated and complex use cases we will be able to support.” The growth of artificial intelligence and machine learning, he adds, makes it vital for organisations to be able to access more data, in order to train their AI and ML systems.
The same principles can be applied in multiple industries, Troncoso points out. The financial services sector, where privacy and security are hugely important considerations, is another major target for the company. “Banks often need to collaborate, but they don’t want to share customer data,” Troncoso says.
Financial crime is one good example – a bank may want to check that a particular account or customer has not been flagged anywhere else, but other providers will want to protect their customer data. Tune Insight’s technology enables one bank to check everyone else’s data without ever needing to identify which bank or account shares their customer, or in what way.
In Switzerland, renowned for its tough laws on data privacy, the company has already worked closely with regulators to win support for its technology. It also provides compliance tools to enable customers to show how the technology causes no problems with the European Union’s general data protection regulation (GDPR). “The potential of our solution for cross-border transfers is recognised by leading legal scholars,” Troncoso adds.
The company is at a relatively early stage in commercialising its software, though it has begun to license its product to a number of customers. “We’re now focusing on internationalisation and scale,” says Troncoso.
Hence today’s fund-raising, with Troncoso particularly pleased to have attracted interest from both US and European backers. The $3.4 million round is led by 4Peaks Capital, with participation from investors including Inflection.xyz, Debiopharm and Zurich Cantonal Bank. Existing investor Wingman Ventures has also contributed.
“Tune Insight stands at the forefront of addressing a critical gap in the realm of confidential collaborative analytics,” believes Edoardo Ermotti, founder and managing partner of 14Peaks Capital. “Its software, situated at the intersection of AI/ML, software-as-a-service, and data security, offers compelling applications across industries, including financial services.”
At Inflection.xyz, founding partner Alexander Lange adds: “AI’s bottlenecks won’t be software development, tooling or algorithms but access to raw compute and high- quality data sets; Tune Insight is making a breakthrough contribution by building an operating system for model training on confidential data without leakage.”
The bottom line, argues Troncoso, is that those hoping to succeed in the data economy need a different mindset. “We should be thinking about how we share insight, rather than data itself,” he says.
Read the full article here