Being a Chinese student in Australia, I have observed that the same experience with TikTok that is rather weird is described in more or less the same words by people around me. You say something in real life like a skincare product, a place to go on a holiday, a political scandal or a celebrity scandal and a few hours later TikTok is showing you videos about the same subject. The initial response is typically instant: Why does TikTok know this already? At other times the question is made more blunt: Is my phone spying on me? I know what reaction is, I have had it. When an application appears to have overheard your last conversation in private, it is difficult not to think that something has gone wrong. However, I do not believe that the most helpful question is whether or not TikTok is literally listening to your private conversations. It is a dramatic question, and it is too specific as well. What is more crucial is that TikTok might not need to listen at all. The same feeling can be delivered by platforms through data collection, behaviour tracking and prediction. This is why it is not simply a tale of a haunting app. It is an online rights matter. The issue of surveillance is not just a matter of surveillance in the basic meaning. That is, it is the fact that the recommendation systems can turn the personal life even less personal even in the cases when no particular noticeable violation is obvious.
In fact, TikTok’s privacy policy explains that the platform is permitted to gather and utilize (to create custom content and services) users’ browsing history, click history, follows, comments, shares, posts, and certain location information (Douyin, n. d. ). They may even alter the results of searches, related searches, and browsing activity based on their location and logs data. The platform is saying that the personalization of the user experience is reliant on data collection and micro-analysis of minor user activity. This doesn’t prove that the suspicious activity recommendations are due to TikTok listening to your conversations. But it does show that TikTok has an extraordinary means of understanding users and predicting their wants. TikTok can learn from users. It can learn from their searches and what they pause on, replay, overlook, their interest, and especially how long they watch. This is why many users may feel like the platform “heard” something, when the system has literally just predicted users’ interests from a large amount of data. It is the epitome of what makes the experience so disconcerting. It is not just the fact that it is accurate, but that it is accurately invisible. Users may not see the underlying reason why they were given a certain recommendation, and that is what makes it unsettling.
What makes this experience disturbing is its near-coincidence and manipulation. It is often difficult for users to distinguish between a recommendation that was generated by a search made by the user, a few seconds of additional viewing, a friend in the same social network, or an inference by the platform based on a broader behavioural pattern. That uncertainty is important, it implies that the issue is not that platforms gather data, but that such data is gathered in the manner that cannot be easily tracked or comprehended by an average user. Whenever individuals believe that TikTok has heard too much about them, they are likely reacting to this insecurities and rather than to any particular evidence. The platform appears to have some personal knowledge, yet the user is not aware of how the knowledge was created. This brings about an unbalanced relationship. One is a predictive, classifying side-watcher. The other party continues to make guesses. What makes the uneasiness about TikTok in this respect is not merely the privacy in the traditional sense of keeping a secret. The loss of intelligibility is also concerned. The user no longer knows what aspects of his life are being monitored and turned into signals, how the signals are being aggregated, or when a casual habit becomes a data point. This is why the sense of being watched is not to be disregarded even in the cases when there is no obvious evidence of literal surveillance. It represents a true asymmetry between what the platforms can learn about users and vice versa (Suzor, 2019; OAIC, 2024).
This is where the study by Chen and Cheung would be particularly helpful. In their article about privacy on WeChat, they refer to a privacy paradox, when users declare that they are concerned abouth privacy, but still remain on the platform and keep providing more information since its benefits are too valuable to lose (Chen and Cheung, 2018). They do not mean that users are irrational. They are saying that, when a platform is so deeply integrated into communication, work, study and everyday life, it becomes not so easy to quit. It turns out to be socially and practically costly. This is a perfect match for TikTok. The majority of the population does not continue using Tik Tok as they do not trust it completely. They continue to use it as it is entertaining, social, fast, and it is entrenched in their daily habits. Chen and Cheung describe privacy using the concepts of accessibility and control as well. The concept of privacy is not simply about the publicly of information. It is also whether users have a choice over what other people know about them, and the extent to which others can know. This is important as platform privacy is now not just about what is explicitly posted. It is also about what platforms can infer from behaviour. The user might never explicitly tell it, I am interested in this topic, but the platform could figure out through watching, pausing, replaying, and searching patterns. That is to say that privacy ceases to be merely about secrets. It is also concerning what can be learned from us without our explicit knowledge. It is not that TikTok is aware of something, in this sense, that makes it uncomfortable. That it can know something personal from behavior which seemed trivial when we created it.
That is why that is commonly used, namely, users have already accepted the privacy policy, is ineffective. All agreements do not have formal meanings. The vast majority of users are unaware of how the data is gathered, how it is processed, and how it becomes the recommendations at Tik Tok. Simply clicking on yes does not fix that issue. It just conceals it in legalese. A system that cannot be viewed or comprehended is expected to be accepted by the users. The argument by Nicolas Suzor comes in handy at this point. In Who Makes the Rules? Suzor contends that platforms are not neutral. They are the creators of rules, they determine what people watch, and how people utilize digital spaces (Suzor, 2019). Users might have a sense that the spaces are theirs since they share on them, discuss there, and develop relationships there. Yet the true strength remains with the platform. This is not a democratic relationship in law. It is a contract, and the contracts are written in such a way that they will do little to actually offer the users bargaining power but are more likely to protect the company (Suzor, 2019). This is because recommendation is not a technical feature. It is part of platform governance. What the user will see is not necessarily what is, but what the platform has chosen to rank, repeat and reward. Neither does it occur at the whim of a product when a feed is too personal. The fact that the platform can regulate attention and visibility to a significant degree and users have little to no idea about the functioning of the power is an indicator of the extent to which the platform has power. Suzor adds that the rule-making on the platforms is also quite often opaque. The way decisions are made, which can often remain invisible to users, fosters suspicion and folk theories when something suspicious happened (Suzor, 2019). That is one of the reasons why users are so fast to leap to the idea that the app is listening. Even though one might not always be technically accurate, it is a logical reaction to a system that obviously responds to users and refuses to answer the questions.
In the eyes of an Australian government, this problem has long since gone beyond individual discomfort or internet paranoia. According to the final report of the ACCC 2025 Digital Platform Services Inquiry, regulators require regulatory reform to deal with competition and harms associated with digital platforms (ACCC, 2025). The OAIC further describes that personal information covers a wide scope of information that may identify the individual, such as IP addresses, photographs, voice and facial biometrics and location data from a mobile device since it may indicate user patterns and habits of use (OAIC, n. d. ). In its guidance on tracking pixels, the OAIC observes that tracking technology can be used to gather data on user activity, that data can be personal, and that covert data gathering without an individual’s consent is most likely to be unfair; it further emphasizes that clear and transparent notice and opt-out should be provided (OAIC, 2024). I would not want to make this a policy essay, as that would move the discussion too far away from lived experience. Nonetheless, these Australian sources are not salient. They demonstrate that issues around platforms are no longer peripheral and overstated. They have become a mainstream topic in consumer harm, privacy, tracking, and accountability. When people feel that TikTok knows them too well, that feeling should not be dismissed as simple fear of technology. It reflects a wider problem. Platforms often know much more about users than users know about platforms, and users have limited control over that gap.
What I find the most interesting is the fact that this feeling can be found in various locations, although people describe it differently. I have also observed as a Chinese student studying in Australia that the words might be different but the pain will remain. In one context, individuals can discuss governance and data control more. In a different one, they can be more concerned about the right to privacy, consumer protection, or big tech trust. The emotion behind it is quite alike, though. It seems to people that the platform is intruding into their lives. They believe that data is being made out of everyday life. They believe that they are exposing personal interests not to another individual but to a system that never ceases learning. This is why I do not believe that the question that would be the most helpful is whether TikTok is good or bad. A more significant question is what type of digital life we are being led to in case recommendation is so central. Privacy is not our responsibility as users, but about the platform design when platforms know us through small indicators, when users themselves are unaware of how their profiles are constructed, and when it is socially costly to quit the platform. In such a case, it is not sufficient to tell the users to be careful. It places excessive burden on the individuals and very little on the systems that generate the issue. A more appropriate reaction would involve seeking clarification of the same, less ambiguous advice, stricter boundaries to the use of the data and more realistic options for users.
The fact that TikTok may have listened to one private conversation is not what is most worrying. What is most worrying is that it may not have needed to. If TikTok can make highly personal predictions about people, the hidden microphone is not the issue. The core issue is that such inferences are integrated into everyday digital life. Privacy is absent in these everyday digital systems. The phrase “TikTok knows me too well” may not be accurate, but speaks to an everyday digital system lacking user comprehension. Recommendation systems are no longer just an interesting feature, but influence attention, behavior, and daily routines. Therefore, it is insufficient to ask whether TikTok is listening. The more important issue is why these systems are allowed to monitor so much user behavior and provide so little explanation. This is, in my opinion, the most pressing issue related to our digital rights.
Reference
Australian Competition and Consumer Commission. (2025). Digital Platform Services Inquiry final report – March 2025. https://www.accc.gov.au/about-us/publications/serial-publications/digital-platform-services-inquiry-2020-25-reports/digital-platform-services-inquiry-final-report-march-2025
Chen, Z. T., & Cheung, M. (2018). Privacy perception and protection on Chinese social media: A case study of WeChat. Ethics and Information Technology, 20(4), 279–289. file:///Users/ruijiechen/Downloads/s10676-018-9480-6.pdf
Douyin. (n.d.). Douyin Privacy Policy. https://www.douyin.com/agreements/?id=6773901168964798477
Office of the Australian Information Commissioner. (2024). Tracking pixels and privacy obligations. https://www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/organisations/tracking-pixels-and-privacy-obligations
Office of the Australian Information Commissioner. (n.d.). What is personal information?. https://www.oaic.gov.au/privacy/your-privacy-rights/your-personal-information/what-is-personal-information
Suzor, N. P. (2019). “Who makes the rules?” In Lawless: The secret rules that govern our digital lives. Cambridge University Press. Book. https://www.cambridge.org/core/books/lawless/8504E4EC8A74E539D701A04D3EE8D8DE
Be the first to comment