What Happens to a Report: Why Women Stop Clicking "Report"

There is one gesture that random video chat users make a lot during their first week, rather less during their first month, and almost never after six months: pressing "report".

It isn't that abusive behaviour goes away. It's that reporting, as it is designed on most random chat services, gives absolutely nothing back to the person who sent it. No acknowledgement of receipt, no follow-up, no confirmation that a human read anything at all. The user clicks, the screen moves on to the next stranger, and the only thing she knows for certain is that she lost two seconds.

We have already described immediate rejection in the first thirty seconds seen from the other side, the permanent exit in invisible attrition, and the bystander's role in the article on the witness. This article deals with a link in the chain nobody looks at: the machinery that kicks in after the click. What it does, what it doesn't do, and why its opacity pushes thousands of users to give up.

What Actually Happens Behind the Button

Three layers, rarely explained

On a random video chat service, moderation generally rests on three stacked levels, none of whose details the user ever sees.

1. Upstream automated detection. Image classification models analyse snapshots of the stream at regular intervals — often one frame every few seconds, sometimes less. They look for nudity, clearly non-compliant scenes, and sometimes faces that appear to be minors. This layer operates before any report: it cuts, blurs or queues. It is fast and crude. It misses most of what users describe: verbal insults, persistent demands, threats — all of which are audio or text, not visual.

2. The user report. Your click produces a record: session ID, timestamp, sometimes one or more frames automatically extracted from the stream at the moment of the report, sometimes nothing more than the IP address of the person reported. The value of that ticket depends entirely on what the platform has kept. On a service that keeps no recordings at all — presented, quite rightly, as a privacy guarantee — a verbal report with no visual evidence is almost nothing.

3. Human review. It exists, but it is sized for volume, not for the individual case. The transparency reports that the major social platforms have been publishing since the European Digital Services Act (DSA) came into force give a useful order of magnitude: millions of reports processed, median handling times counted in hours for widely viewed content, and high proportions of decisions taken without individual human review. Random chat services, which are far smaller, rarely publish the equivalent.

The ephemerality paradox

This is the heart of the problem, and it is structural rather than malicious.

Random video chat is built on the absence of any trace: no sign-up, no history, WebRTC streams that are peer-to-peer or relayed but never archived. That architecture genuinely protects users — we explained it in our article on screenshots and AI nudification: the real risk comes from the person on the other end, not from the platform.

But the same architecture makes a complaint hard to investigate:

| What the user experienced | What the platform can reconstruct | | --- | --- | | A few seconds of indecent exposure | One automatic frame, if the capture happened at the right moment | | Sexist insults spoken out loud | Nothing, unless audio is analysed in real time | | Persistent demands via text chat | The messages, often kept for a few minutes | | Threats after a refusal | Nothing, the session is already closed |

In three cases out of four, the user is the only witness, and her testimony is recorded nowhere — because the report form almost never asks her to describe anything at all. One button, three categories, submit.

Why the Lack of Feedback Destroys Trust

A well-documented mechanism

In behavioural psychology, reinforcement is what keeps an action going over time. A repeated action that produces no perceptible feedback dies out. That is exactly what happens to reporting: not because users doubt its moral usefulness, but because nothing, ever, confirms an effect to them.

French research on cybersexism points the same way. The study conducted by the Centre Hubertine Auclert on cybersexism among teenagers, and more broadly the Défenseur des droits surveys on non-take-up of rights, describe the same sequence: the victim judges that the process will be long, pointless, or even turned against her, and gives up. The volume of real incidents is therefore always far higher than the volume of reports. A platform's moderation figures do not measure harassment: they measure the share of harassment that people still had the energy to declare.

The hidden cost of reporting

Pressing a button seems free. For the person doing it, it isn't.

• You have to stay. Reporting means not skipping immediately, so spending a few more seconds looking at what you wanted to escape.
• You have to classify. Choosing between "nudity", "harassment" and "minor" forces you to name the incident while you are still in it.
• Sometimes you have to prove it. Forms that ask for a screenshot require the victim to keep the very image she wants to forget.
• You have to start over. A report does not protect the next woman in the queue that same minute.

The result: silent skipping becomes the rational behaviour. Faster, less costly, apparently just as effective. And the platform, receiving nothing, concludes that everything is fine.

What European Law Now Requires

Regulation (EU) 2022/2065 on digital services, the DSA, applicable to all online platforms in the Union since February 2024, changes things on paper. It does not solve everything, but it creates concrete obligations that users can rely on.

What the text requires of any platform hosting user content:

• an accessible, easy-to-use notification mechanism (Article 16), which must allow the report to be substantiated, not just a box to be ticked;
• an acknowledgement of receipt and information about the decision taken, together with its reasons;
• a statement of reasons sent to the user whose content or account is restricted (Article 17);
• an internal complaint-handling system, free of charge, to challenge a decision (Article 20);
• the option of referring the matter to a certified out-of-court dispute settlement body (Article 21);
• notification of the authorities where there is a suspicion of a criminal offence threatening a person's life or safety (Article 18);
• measures against abusive reports and repeated manifestly illegal content (Article 23).

In France, Arcom is the designated Digital Services Coordinator for this regulation. In practice, when a platform accessible from the Union offers no acknowledgement of receipt and no complaint route, it is not merely unpleasant: it is in breach.

The three remedies that really exist

A user confronted with forced exposure or online sexual harassment is not limited to the in-app button. We set out the criminal law framework country by country in our article on cyberflashing and the laws in France, Spain, Germany and Brazil. On the reporting side, three channels coexist:

1. The platform, via its DSA notification mechanism — to be used by describing the facts in writing, however briefly.
2. The official State reporting platform. In France, the PHAROS system (internet-signalement.gouv.fr), run by the anti-cybercrime office, receives reports of illegal content and behaviour. 17Cyber, launched in late 2024 by the Ministry of the Interior, provides guidance and connects users with the relevant service. The e-Enfance / 3018 association supports minors and their families.
3. Filing a complaint, at a police station, a gendarmerie, or by letter to the public prosecutor. Online complaints are also possible for certain digital offences via the dedicated public service.

An internal report and a criminal complaint are not mutually exclusive. The first can get access cut within hours; only the second can lead to judicial proceedings.

How to Report So That It Actually Counts

If you decide to report, the ticket may as well be usable. A few habits noticeably change the value of a case.

Pin down what can be pinned down, without putting yourself at risk

The decisive element is almost always the timestamp. A platform that keeps no streams very often keeps connection logs. Saying "24 October at 10:07 p.m., session via Firefox browser, desktop computer" makes it possible to trace a pairing; saying "last night some guy showed me his bits" makes nothing possible.

Systematically note down, in a small paper notebook kept beside the keyboard or in a phone note: date, time to the minute, nickname or displayed ID if there is one, country shown, and the nature of the events in one sentence. Those thirty seconds turn a feeling into a verifiable element.

For serious cases, a screenshot may be necessary — but it must stay under your control. The useful habit is to store it outside your personal photo stream, for example on an encrypted USB stick reserved for that purpose, and never to share it again. In France, keeping evidence for the purpose of filing a complaint is permitted; republishing it, on the other hand, can expose you.

The form: write something, even two lines

When a free-text field exists, use it. A substantiated report leaves the automated queue and more often lands in front of a human. Three elements are enough: what happened, when, and what you are asking for (access cut off, evidence preserved).

Protecting the session itself

Many incidents end badly because the user is caught off guard: camera on, face identifiable, room recognisable. Two very ordinary accessories reduce the cost of a bad exchange. A sliding webcam cover lets you cut the image instantly without hunting for a software button. A headset with a noise-cancelling microphone means you don't have to speak loudly and keeps the room out of the audio frame. These are not security tools in the strict sense, but they make leaving a conversation far quicker.

What Men Can Actually Do

Reporting is presented as a matter for victims. That is an arithmetical error: on a random video chat, the male population is overwhelmingly in the majority — we examined the available data and its limits in our analysis of the gender imbalance. The reports that carry the most weight are therefore, potentially, the ones coming from them.

Three actions, no heroism required:

• Report what you come across, even if it wasn't aimed at you. An exposure stream runs on a loop and meets dozens of people an hour. One ticket per fifty encounters is detection by chance; fifteen tickets in ten minutes is an automatic priority.
• Report fake profiles and looped videos. They are often the gateway to paid-link scams, described in our guide to spotting looped videos and paid links in 30 seconds.
• Don't confuse disagreement with abuse. Reporting someone who simply skipped you clogs the queues and degrades the handling of real cases. The DSA explicitly provides for measures against abusive notifications.

What an Honest Platform Should Display

To finish with a usable checklist. A video chat service that takes the question seriously makes five things visible — and their absence is, in itself, information.

1. An acknowledgement of receipt after every report, with a ticket number.
2. A stated processing time, even an approximate one.
3. Information about the decision, including when it is negative.
4. A free complaint route to challenge it.
5. A periodic transparency report: volume of reports, share handled automatically, measures taken.

None of these five points is technically difficult. Their absence is a choice, not a constraint.

To understand how the tools available to you during a session actually work, our article on the real-world use of moderation tools sets out the difference between blocking, reporting and leaving. And if you want to try a conversation under better conditions, the simplest approach is still to jump into the chat with a clear idea of what you'll do if things go wrong.

A report is not a vote of confidence sent into the void. It is data. And in a system where victims have stopped producing it, the only data circulating is the data saying that everything is fine.

---

Main sources: Regulation (EU) 2022/2065 on digital services (Articles 16 to 23); Arcom, Digital Services Coordinator for France; PHAROS platform (internet-signalement.gouv.fr), Office anti-cybercriminalité; 17Cyber system, Ministry of the Interior; Centre Hubertine Auclert, work on cybersexism; Défenseur des droits, reports on non-take-up of rights.