A software and hardware that guarantees privacy isn’t compromised by the vast quantity of data required to train Artificial intelligence (AI) models, was offered by the Taiwanese startup DeCloak.
Nowadays, the AI is everywhere, so, many innovations linked to that idea. As the IT-experts say, there is a problem at the heart of AI innovation – those models require huge amounts of data, and many worry about what this means for privacy.
DeCloak is the Taiwanese startup, which has offered software and hardware that addresses this challenge. It could help businesses to strike a balance between the potential benefits of AI and the need to preserve privacy rights of the customers.
In general, de-identification is a process that converts personal data (any information related to an identifiable person) into non-personal data so that it can be used by third parties while also preserving privacy. This is particularly important for any big data or data fusion applications.
DeCloak provides de-identification in the form of three pieces of hardware that can be plugged into a range of everyday gadgets. The company’s privacy processing unit (PPU) is designed like an Application-specific Integrated Circuit (ASIC) – a type of chip that is specialised to perform a particular role and is compatible with various internet of things (IoT) devices.
The Taiwanese firm also provides a USB dongle that works with devices that have a USB adapter, and a SIM card that works with smartphones. When confidential data needs to be transmitted, these plug-ins directly de-identify the data via an embedded algorithm. The company also provides a range of software packages that perform a similar role. One of the key benefits of DeCloak’s technology is that de-identification happens in real time.
The industries that could particularly benefit from de-identification are those that hold large amounts of personal data that could be used to personalise and optimise services – notably banking and healthcare. Logistics is another industry that DeCloak is targeting. To ensure that customers receive their packages, some data must be exchanged between e-commerce platforms and logistics firms.
But e-commerce companies are reticent about handing over personal data and logistics companies are themselves cautious about the reputational and legal risks inherent in holding such information. If logistics services could be carried out using de-identified data, it would be a win-win all round.
