The Specific Event

Apple updated its App Store Review Guidelines this week to more explicitly prohibit applications that generate intimate images of people without their consent, including what the guidelines now specifically call "bait-and-switch" apps that advertise one function but contain hidden "nudification" features. This is not Apple's first attempt to address this category of harm. The revision signals something organizationally interesting: a platform that is iterating its ruleset in response to demonstrated regulatory failure. The previous rules were apparently insufficient to stop the behavior, so Apple wrote more specific rules. This is the part that deserves scrutiny.

When Specificity Is the Wrong Response to a Structural Problem

The instinct behind Apple's update is understandable. A harmful category of application exists. The existing rules did not clearly prohibit it. Therefore, clarify the rules. This is procedural governance logic, and it is the dominant mode of platform policy design. The problem is that this approach treats the App Store as if it were a regulatory code where sufficient specificity produces compliance. Platform environments do not work that way.

Kellogg, Valentine, and Christin (2020) documented in detail how workers and actors operating within algorithmically-mediated environments do not respond to rules as written. They respond to the enforcement pattern they can observe. The "bait-and-switch" structure that Apple's updated guidelines now name explicitly exists precisely because developers learned, through repeated interaction with the review process, where the detection threshold sat. Adding a new named category to the ruleset does not change the underlying dynamic. It shifts the location of the threshold.

The Awareness-Capability Gap, Applied to Governance

My dissertation research focuses on what I call the Awareness-Capability Gap: the empirical finding that algorithmic literacy research consistently shows workers develop awareness of how algorithms operate without that awareness translating into improved outcomes (Kellogg et al., 2020; Gagrain, Naab, and Grub, 2024). I have been applying this framework to platform workers, but Apple's nudify ban reveals the same structure operating at the governance level.

Apple is clearly aware that the problem exists. The updated guidelines demonstrate sophisticated awareness of specific evasion tactics, naming the bait-and-switch pattern by name. But awareness of the evasion pattern is not the same as the capability to prevent it at scale. The App Store processes millions of submissions. The review infrastructure that catches explicit policy violations operates on a combination of automated screening and human review. Adding conceptual categories to a guideline document does not obviously improve either of those mechanisms. The awareness-capability gap applies to the platform operator, not just to the platform participant.

What Topology Has to Do With It

A distinction I use in my theoretical work separates topology from topography in platform contexts. Topography is the specific landscape of rules, procedures, and checkpoints as they exist at a given moment. Topology is the underlying structural shape of the constraint environment, which persists across changes to the topography. Apple's update changes the topography. It adds an explicit node to the map of prohibited content. But the topology of the App Store governance problem remains unchanged: a large-volume, asymmetric review process where the reviewer cannot fully inspect application behavior and the developer controls the submission framing.

This topological structure is what produces bait-and-switch behavior in the first place. It is not a product of vague guidelines. It is a product of information asymmetry between the reviewing party and the submitting party, which is a structural feature of the review environment. Rahman (2021) described a similar dynamic in gig platform governance, where the "invisible cage" of algorithmic control shapes behavior not through explicit rules but through structural constraints that persist regardless of how those rules are worded.

What Effective Platform Governance Would Have to Look Like

If the analysis above is correct, Apple's revised guidelines will produce a temporary reduction in easily detectable nudify applications and a corresponding adaptation in evasion tactics. This is the standard iterated-rule-specificity outcome. Genuine structural governance would require addressing the information asymmetry directly, through dynamic behavioral monitoring post-approval rather than static pre-approval review, or through developer accountability mechanisms that survive the submission checkpoint.

Neither of those approaches is simple. Both carry significant resource and false-positive tradeoffs. The point is not that Apple has an obvious solution available and chose the wrong one. The point is that organizations under public pressure tend to reach for the tool that produces visible, documentable action: the updated policy, the renamed category, the new guideline. The structural problem that generated the behavior remains, and it will generate the next variant. Platform governance that conflates topographic revision with topological change will keep arriving at this same juncture.

References

Gagrain, A., Naab, T. K., and Grub, J. (2024). Algorithmic media use and algorithm literacy. New Media and Society.

Kellogg, K. C., Valentine, M. A., and Christin, A. (2020). Algorithms at work: The new contested terrain of control. Academy of Management Annals, 14(1), 366-410.

Rahman, H. A. (2021). The invisible cage: Workers' reactivity to opaque algorithmic evaluations. Administrative Science Quarterly, 66(4), 945-988.