Skip to main content

Open for Debate

Unravelling the Complexity of Manipulation Theories

15 April 2024

In the previous four posts of this series on Unpacking Manipulation in the Digital Age(Post 1; Post 2; Post 3; Post 4), I argued that more attention to different types of social influence is needed and that digital influence often seems to fall into the grey area between persuasion and coercion.

This raises the question of whether and why they fit the category of manipulation. In this post, I argue that, in the pursuit of understanding manipulation in the age of new technologies, two prevalent manipulation theories are available, each with its own set of problems.

Intention-Focused Theories

Philosophical and legal scholars often emphasise intention in theories of manipulation (Sher 2011; Spencer 2020; Noggle 1996). Intentions, they argue, are crucial for distinguishing between purposeful influence and accidental outcomes. Robert Noggle’s manipulation theory stands out, suggesting that manipulators intend to make their victims commit a mistake (Noggle 2020).

However, this approach falls short in the realm of digital influence, where the intent behind a given influence is often elusive. One reason is that we are dealing with AI-mediated influence (see blog post 3 ). When a particular influence – like a blog post or an image posted on Facebook – is mediated by AI, intentions are either missing or hard to identify. And yet misleading or otherwise problematic online influences may strike us as manipulative even if they lack the typically nefarious intentions required by intention-focused theories.

Another reason is that some of the influences that occupy the grey area between persuasion and coercion may be completely bereft of intention. A/B tests are a good example. A/B tests are a user experience research method, which usually involves two variants of a design (A and B) that are randomly distributed to users. In that way the designer can test, for example, whether variant A of a website design works better in terms of e.g. increasing sales on the website. Brignull (2023) describes how A/B-tests can be completely automated to evaluate which, for example, website design works most effectively in generating traffic and to implement the winning design automatically. An unsuspecting designer may simply intend to create a website design vaguely defined as good and end up, by way of the automated process, implementing a design that uses exploitative but effective dark patterns. This will strike many as a manipulative form of influence, and yet, the process, however, lacks the nefarious intentions that intention-based theories require.

In addition to this general problem for intention-based theories, there are significant obstacles for specific interpretations of intention-based manipulation theories. Perhaps the most influential one is the covertness view of manipulation formulated by Daniel Susser, Beate Roessler, and Helen Nissenbaum (Susser et al. 2019a, 2019b). The problem is that it develops an overly stringent necessary requirement on manipulation – that it be covert – which, as others have pointed out, is often not present in clear cases of manipulation (Bongard-Blanchy et al. 2021; Noggle 2022; Klenk 2022).

Disjunctive Characteristics Theories

The digital ethics literature presents theories that consider manipulation as a disjunction of different characteristics. This laundry list approach lists various factors associated with manipulation, such as undermining autonomy or causing harm, and suggests that manipulation is a type of influence that satisfies some or all of these criteria (Botes 2023; Ienca 2023). The challenge arises when trying to find common ground among these diverse features. In particular, disjunctive views struggle to provide a clear understanding of manipulation and pose practical challenges in formulating effective design or policy recommendations, raising both practical and theoretical challenges.

The Practical Challenge for Disjunctive Views

Disjunctive views make it challenging to pinpoint specific measures against manipulation. For instance, an effort to avoid manipulation by being transparent (stipulating that ‘covert influence’ is one of the disjuncts of a manipulation criterion) may conflict with the goal of preventing harm, creating a practical dilemma. This complexity hinders the application of concrete solutions in design and policy recommendations (Klenk 2023).

More generally, an account of manipulation should eventually allow us to derive concrete design recommendations to implement regulatory measures against manipulation in practice. A long list of features merely indicative of manipulation may complicate the task of deriving consistent design requirements too much.

Therefore, unless we are sufficiently convinced – which we should not be at the moment – that there is no more straightforward, unified criterion of manipulation, we should avoid disjunctive views and keep looking for a unified criterion.

The Theoretical Challenge for Disjunctive Views

Disjunctive views also leave a theoretical lacuna. Unless we can find a commonality or unifying factor behind the different disjuncts allegedly related to manipulation, we fail to truly understand the phenomenon.

One aspect of this theoretical lacuna is that we are missing an answer to the question of what, if anything, seemingly widely different forms of manipulation have in common. After all, one instance of manipulation may fit a subset of the disjuncts that another instance of manipulation does not share. What then warrants us classifying both these influences as part of the same phenomenon (Klenk 2024)?

Another aspect of the theoretical lacuna is moral. As Coons and Weber (2014) point out, it seems that there is a fairly uniform moral response to manipulation. It appears to be a problematic form of influence, and various scholars are trying to pinpoint precisely why manipulation is supposed to be a moral problem (Klenk and Hancock 2019). With a disjunctive view, however, it is no longer clear that we can assume that all forms of manipulation should receive the same moral response on a basic level since, as discussed, different forms of manipulation may have nothing in common. When manipulation is used as a meaningful category in policy and regulation, warranting bans and punishment, then this becomes a real practical and legal issue, too.

Conclusion and Outlook

The issues for two prominent approaches to understanding manipulation outlined in this post result in a call for a better theory.

In this series’s sixth and final post, I‘ll review a potential avenue for a more accurate and actionable manipulation theory: the indifference account of manipulation.


Bongard-Blanchy, K., Rossi, A., Rivas, S., Doublet, S., Koenig, V., & Lenzini, G. (2021). I am Definitely Manipulated, Even When I am Aware of it. Its Ridiculous! – Dark Patterns from the End-User Perspective. In W. Ju, L. Oehlberg, S. Follmer, S. Fox, & S. Kuznetsov (Eds.), DIS ’21: Designing Interactive Systems Conference 2021, Virtual Event USA, 28 06 2021 02 07 2021 (pp. 763–776, ACM Conferences). New York,NY,United States: Association for Computing Machinery. doi:10.1145/3461778.3462086.

Botes, M. (2023). Autonomy and the social dilemma of online manipulative behavior. AI and Ethics, 3, 315–323. doi:10.1007/s43681-022-00157-5.

Brignull, H. (2023). Deceptive Patterns: Exposing the tricks tech companies use to control you. Harry Brignull.

Coons, C., & Weber, M. (2014). Manipulation: Investigating the core concept and its moral status. In C. Coons & M. Weber (Eds.), Manipulation: Theory and practice (pp. 1–16). Oxford: Oxford University Press.

Ienca, M. (2023). On Artificial Intelligence and Manipulation. Topoi, 42, 833–842. doi:10.1007/s11245-023-09940-3.

Klenk, M. (2022). (Online) Manipulation: Sometimes Hidden, Always Careless. Review of Social Economy, 80, 85–105. doi:10.1080/00346764.2021.1894350.

Klenk, M. (2023). Algorithmic Transparency and Manipulation. Philosophy & Technology, 36, 1–20. doi:10.1007/s13347-023-00678-9.

Klenk, M. (2024). Ethics of generative AI and manipulation: a design-oriented research agenda. Ethics and Information Technology, 26, 1–15. doi:10.1007/s10676-024-09745-x.

Klenk, M., & Hancock, J. (2019). Autonomy and online manipulation. Internet Policy Review.

Noggle, R. (1996). Manipulative Actions: A Conceptual and Moral Analysis. American Philosophical Quarterly, 33(1), 43–55.

Noggle, R. (2020). Pressure, Trickery, and a unified account of manipulation. American Philosophical Quarterly, 57, 241–252. doi:10.2307/48574436.

Noggle, R. (2022). The Ethics of Manipulation. In E. N. Zalta (Ed.), Stanford Encyclopediaof Philosophy. Summer 2022 .

Sher, S. (2011). A Framework for Assessing Immorally Manipulative Marketing Tactics.Journal of Business Ethics, 102, 97–118. doi:10.1007/s10551-011-0802-4.

Spencer, S. B. (2020). The Problem of Online Manipulation. University of Illinois Law Review, 2020, 959–1006. doi:10.2139/ssrn.3341653.

Susser, D., Roessler, B., & Nissenbaum, H. (2019a). Online Manipulation: Hidden Influences in A Digital World. Georgetown Law Technology Review, 4(1), 1–45.

Susser, D., Roessler, B., & Nissenbaum, H. (2019b). Technology, autonomy, and manipulation. Internet Policy Review, 8, 1–22. doi:10.14763/2019.2.1410.

Image by <a href=”″>Robinraj Premchand</a> from <a href=”″>Pixabay</a>