Privacy Concerns in AI

Note: The article in copy is present on authors medium account as well.

Here in this article, we look at some of the concerns that occur when using AI. Yes, it is Privacy Issues with AI we shall cover in this article. The topics we shall cover range from AI-based data generation to data collection to data usage, data storage and data access and re-access.

As per Bartneck, 2021 [1] privacy is right so that no one can observe you. Privacy definition can be extended to respect the non-declaration of preferences, choices, interests, and wants a person has.

Here are some key points in the lifecycle of data where AI can lead to privacy issues.

1. Data Generation

A lot of data of each internet user is generated hourly. This data once generated can be misused for a variety of purposes such as personalized recommendations to things to shop, to even AC temperature ones like, the Sunday match updates. Even the most intimidating wishes are collected (Bartneck, 2021) [1]. All these need to be stopped so that personal data does not go to the wrong place. Some measures such as “Do not track” should be set on the browser. There are other ways too when legitimate data is generated. This can be on local stand-a-lone devices or connected devices such as gadgets, apps, and or even gym machines. Examples of legitimate data collection are hourly weather data generation or hourly space data collection where in terabytes of data is generated regularly.

2. Data Collection

Data collection examples are like drones collecting data while out for the delivery of medicine, packets, or even for surveillance. Each day huge amount of data can be collected. There are other legal ways of data collection which include agents for data collection, data collection by asking opinions from users in the form of form filling, other things including surveys. All these data are legitimately collected if users allow the use of data and have given their consent.

3. Data Usage

Data can be collected, stored, or sold for some reason. This collected data can be used to alter the decision of people for voting result manipulation or to buy a particular company’s products. AI can impersonate users with this large data about people. Some other kinds of data are those generated by medics and the entire profession of medicine. Here, data generated in surgery or so can be used for not only future suggestions but also advice to patients with similar symptoms. Insurance companies may collect data or buy data to predict with AI the risk to a person and risks of people with similar stats. Another use of data is to compute how good a citizen is with AI, speculations show China uses such AI algorithms for such mass surveillance.

4. Data Storage

Another concern is while interacting on the internet the system may be subject to hacking, which may leak many parts of data which was private to the user. Hence, hacking and the privacy of data are interlinked too. Data should be stored in a more secure configuration. All this so that hacking does not compromise privacy issues of data.

5. Data access

Data is like a goldmine to so many companies for it can help in personalized targets, marketing, advertisements (Bartneck, 2021) [1] and hence the sale of products. One tech giant in 2018 was asked how is his business running when it has no way to generate revenue from people. The businessman says they run advertisements on their apps. Well, this is just one such example of how and when data is required to be accessed. All this lead to privacy issues for all this data that business companies and some governments want belong to people, who do not want machines, software, and other humans to know.

6. Later data usage

Data may be persistently required in many cases for example a personal robot need to look up in database several time to know specific preferences of owner. And in the same way, other Smart AI devices may also need a lookup of users’ data. It must be noted accessing data is allowed if the user using it finds it okay, and as long as it serves its purpose of accurate results a user needs. Privacy concerns comes into the picture when these devices/Robots/gadgets contact the server for some information. While doing so, it may intentionally or unintentionally lose/pass some information to a server.

However, privacy can be sacrificed by a good number of people if people see the benefit of the same surpassing the negative side. Examples of these include mass surveillance to reduce crimes. The use of drones is another issue where data is generated, collected, analysed, and even monitored for various uses. All this needs to be explained as to why collecting such data is important for people to allow it.

On the other hand, a user should hold the right to privacy concerns related to the user data. As each day passes the data generated and collected is increasing and hence the need, for awareness and caution to privacy policies of data used bt AI. More user-friendly consent forms should come in to solve the problem of privacy grants and issues. There are many more privacy issues which shall be dealt with coming times.

References

[1] C. Bartneck et al., (2021) An Introduction to Ethics in Robotics and AI, SpringerBriefs in Ethics, https://doi.org/10.1007/978-3-030-51110-4_8