PET technologies or how to give away your data without giving them away

Nowadays, many companies require information and data exchange as a core part of their activities. This poses a problem, as many of us are concerned about the privacy and security of our personal data. Let's take a closer look at what PET Technologies are and how they can address this issue.

WHAT IS PET TECHNOLOGY?

Nowadays, many companies require information and data exchange as a core part of their operations. In fact, some companies' main profit comes from collecting information and then using it to their advantage or selling it to third parties, mainly for commercial purposes. 'Big Data' is here to stay, and it's the foundation of many future — and almost magical — projects like AI and the acclaimed GPT.
But this raises a problem, as many of us are concerned about the privacy and security of our personal data, and more and more people are choosing to stop using certain platforms or online services in favor of protecting our privacy.
This is where these technologies come into play. PET technologies, short for Privacy-Enhancing Technologies, are a set of cryptographic techniques and tools that, based on mathematical equations, offer solutions to share personal — and possibly sensitive — information with third parties without putting our privacy and security at risk.
In other words, these technologies aim to allow us to share our data to gain technological benefits “without actually sharing our data.” While it may sound like a trick out of Hogwarts, these technologies first appeared back in the 1980s, although it's only now that they are becoming increasingly relevant.
Based on two assumptions, we find two types of PET:

SOFT PRIVACY TECHNOLOGIES

Soft Privacy Technologies are based on the assumption that we can trust the company or entity managing the data. Some examples of these technologies include Differential Privacy (DP) and even the SSL/TLS protocol.

HARD PRIVACY TECHNOLOGIES

Hard Privacy Technologies, on the other hand, are based on the assumption that we cannot trust the company or entity managing the data, so no entity should be able to violate data privacy. An example of this would be the use of VPNs or onion routing to perform a task in which you participate personally, or the “secret ballot”, Australia's fully anonymous voting system.

WHY ARE THEY IMPORTANT?

Technology and its advancement, not only in the computing field with the well-known Moore's Law, has an exponential growth rate, which is especially difficult for the human brain to grasp due to its very structure. Let's remember that the first personal computer was built in the 1970s and today, 2/3 of the world's population already carries a smartphone in their pocket, with far greater capabilities than those early computers.
Due to this acceleration and our increasingly digital world where personal, financial, and sensitive information is exchanged, it seems clear that we need an infrastructure where all of this remains possible, without it becoming a threat to us or our privacy. This is especially relevant in areas such as financial institutions, healthcare, elections, and government-related tasks involving citizens.
In addition, PET technologies could arrive to solve all these situations and risks dominated by identity theft, fraud, and the violation of our data—issues that are increasingly present in our digital lives, where AI systems also need our data to keep training and improving.

HOW DOES IT WORK?

PET technologies work through the use of advanced cryptographic and computational techniques that help maintain and protect the privacy and security of the information being shared, especially online. There is no standardized classification, but we can define different types of PETs and their roles or functionalities.

PROTECTION OF DATA

These are techniques that aim to protect user data. Among them, we find techniques such as:

Homomorphic encryption: allows calculations to be performed on encrypted data without the need to decrypt it, ensuring data privacy. It has limitations regarding the volume of data.

Zero-knowledge Proofs (ZKP): allows the verification of a statement’s truth without revealing the actual information or any additional data.

Differential computation: enables calculations on private data without disclosing additional information by adding noise to the data before performing calculations.

Private information retrieval: allows accurate information to be obtained in a fully private or anonymous manner.

DE-IDENTIFY OR ANONYMIZE DATA

In this case, these techniques aim to eliminate and thus anonymize any identifying reference. We find techniques such as:

Differential privacy: information is obfuscated by generating or adding random noise in a way that allows data analysis without revealing individual information.

Synthetic data: data is artificially generated with the same structure and content as real data, but it does not correspond to any specific or real individuals.

Anonymization: techniques where personal information is removed so that individuals cannot be identified.

Pseudonymization: techniques where personal information is partially removed in a way that does not fully compromise user privacy.

COLLABORATIVE ANALYSIS

Lastly, those that aim to enable collaboration without revealing the actual data. We find techniques such as:

Federated analysis: techniques where each person contributes their data, which is processed locally and then combined, preserving each individual's privacy.

Multi-party computation: techniques where data is split into multiple parts, each processed individually. The results are then shared without disclosing the original data.

NEW OPPORTUNITIES AND UNEXPLORED PATHS.

Undoubtedly, PET techniques will play a decisive role in the coming years and decades, as people become increasingly aware of what we share online and its consequences, and where the privacy and security of our data must prevail.
Moreover, as we’ve seen earlier, computing is becoming more efficient and powerful, paving the way for increasingly complex and optimized calculations that support methods where user privacy and cybersecurity are the main priority.
Another area where we may see advances and unexplored opportunities is in the health sector, enabling broader studies with more complete population samples and increasingly accurate results, always ensuring the security and privacy of the data.
All of this will evidently bring about new regulations and legislation to protect data, where I’m sure we’ll see progress at a national and even global level in this regard.

CONCLUSIONS

These technologies may seem like magic, taken from a utopian future, although they are very much part of the present and are already starting to gain momentum. The technological advancements we’ve witnessed in recent years are truly incredible, but we must not forget that many of these breakthroughs are fueled by our data. That’s why we must start becoming aware of our data—who we share it with, what will be done with it, and, in any case, take control over it and decide for ourselves whether or not we want to protect our privacy. We will remain attentive to the future and to this type of technology and its developments, and observe how they take shape and are implemented in society.