1996: a joke among psychiatrists
In July 1996, on a forum about the psychology of the internet frequented by American psychiatrists, a doctor posted a proposal for an "internet addiction support group", complete with a diagnostic grid modelled on the American reference manual. The page's address contained the word humor. It was a hoax, and its author later said so.
The grid was funny to anyone reading it properly. It described a withdrawal syndrome featuring psychomotor agitation, anxiety, obsessive thoughts — and "voluntary or involuntary typing movements of the fingers", as though someone deprived of the internet kept typing in mid-air. For the rest, you only had to replace "pathological gambling" with "internet" in the existing classification: the parody lay entirely in that copy-and-paste.
The forum was widely read. Some readers took the grid seriously. Within months the concept had spread across the internet and into the press, without a single study having established it.
The adoption, and the conflict of interest
The following year an American psychologist turned it into a ten-item questionnaire, presented at a professional conference. The items track the hoax point by point: constant preoccupation, needing more time online for the same satisfaction, inability to control the time spent, irritability on disconnecting, lying to family, risking work or studies. Four positive answers were enough to classify someone as dependent.
Three weaknesses, rarely mentioned at the time, are obvious now:
- The sample. The founding study was run online, among volunteers recruited on the internet. People who answer a questionnaire about internet addiction, on the internet, are not a sample of the population: this is selection bias in its purest form.
- The criteria. They were borrowed from another disorder. Finding that they "work" proves nothing: a sufficiently general grid always classifies somebody.
- The conflict of interest. The author of the questionnaire also ran a paid internet addiction treatment service. Whoever defines the illness should not be selling the cure.
What the survey did show
One finding from that first study has held up: the problem was never "the internet" but particular uses. Only a small minority of respondents described themselves as simply browsing; most fell into chat spaces and networked games. Looking things up appeared nowhere. Twenty-five years later, that is exactly where the real progress has been made.
Where medicine actually settled
The debate did not end by recognising "internet addiction". It ended by recognising, far more narrowly, a disorder tied to one specific use:
| International Classification of Diseases (WHO) | Gaming disorder is a diagnosis in its own right in the eleventh revision, in force since 2022. Three criteria over twelve months: impaired control over gaming, increasing priority given to gaming over other activities, and continuation despite negative consequences — with significant impairment of functioning. |
|---|---|
| American manual (DSM-5) | Internet gaming disorder appears in the section for "conditions for further study": a holding status, not an established diagnosis. |
| "Internet addiction" | Appears in neither. |
The gap between the two grids is measurable: applied to the same population they identify neither the same people nor the same proportions — prevalence estimates roughly double or halve depending on which criteria are used. That is the mark of an object still poorly stabilised, not of an epidemic.
The useful question is not "am I addicted?"
Trying to fit yourself into a box is of little use, all the more so when the box does not exist. The useful question is one of displacement: what is this use taking the place of? Sleep, meals, work or study, real relationships, other pastimes. When the answer is "nothing", there is no problem, whatever the hour count. When it is "sleep, for six months", there is one — and it deserves addressing on its own terms.
The signals that warrant talking to a professional are the same as for any engulfing behaviour: loss of control despite wanting to stop, continuation despite concrete consequences, and distress. See screens and digital use and mental health.
The lesson in method
The episode has become a teaching case in its own right. It shows how a diagnostic category can be born without data, spread through the media rather than through research, be stabilised by those with an interest in its spread, and outlive its refutation by decades. The reflex it recommends goes well beyond the internet: ask who produced the criterion, on what sample, and what that person is selling.