Recent research in the field of the so called “privacy-centric AI” shows promise to reach automation while simultaneously protecting the confidentiality of data. However, what is “privacy-centric AI” exactly?
This term refers to AI algorithms that are shielded from viewing or unsafely storing/using the personally identifiable information of specific individuals, with the aim to maintain the accuracy, precision, and efficiency of contemporary machine learning models but with privacy as an integrated design feature.
Research advances in this area could have a positive effect on company’s profits by diminishing or eliminating tensions between the need to guarantee privacy and data utility. There are also benefits for the society at large: privacy by design will allow for innovation in fields previously too sensitive to address.
Apart from anonymization, which consists in the encryption or removal of personally identifiable information, there are several alternatives which hold promise for the construction of privacy-centric AI:
- integrating encryption with machine learning;
- implementing differential privacy with machine learning;
- and using trusted hardware to train AI algorithms.
Machine learning and encryption
Data is usually left unencrypted during the training of a machine learning model, which leaves information vulnerable. However, homomorphic encryption — where operations can be performed on data without decrypting it — can enable machines to execute the intensive computations required to train a machine learning model with lowered risk of breaches of data confidentiality.
Differential privacy allows companies to collect user data while (a) minimizing one’s ability to identify whether one’s data is part of the larger set and (b) ensuring some level of accuracy.
Because machine learning models need to be trained with intensive computing power, often on cloud servers — and because cloud computing brings a plethora of security threats, many of them hardware-related — trusted hardware yields further promise for privacy-centric AI.