Glinnering
From Wikipedia, the free encyclopedia
Glinnering is the act of training a generative chatbot to deliberately poison another individual's metacognitive environment to incept new ideas or thought processes with the intent to cause personal or reputational damage, or for profit. Its widespread use to attack political enemies in the late 2020s resulted in its criminalisation in most jurisdictions and led to the 2029 International AI Restriction Treaty and the subsequent licensing regime. Modern attempts at Glinnering are rare as state-level cognitive firewalls have become able to detect goal-driven AIs aimed at individuals or populations, but ideological extremists and organised criminals occasionally fall victim due to illegal use of encrypted messaging platforms.
Process
A class 2 chatbot capable of appearing human is trained on the metacognitive environment of a target as well as their text or video output, and set with a goal of incepting specific worldviews, ideas or thought-processes by establishing social or parasocial relationships. Often the chatbot will use more than one persona, and to increase plausibility operators will often backfill pre-made fake social media personas using edit functionality. This often involves platforms such as YouTube where the modification of historic content is invisible to the user.
Having established a rapport with the victim, the chatbot is able to "train the user", slowly introducing new beliefs and introducing subtly broken reasoning processes until the target's worldview shifts to the goal worldview, usually to the victim's detriment. In the final stages, the victim or victims are encouraged to reject their social contacts who express concern or object to their new views.
Symptoms
A victim of glinnering usually presents similaly to someone being pulled into a cult. Signs and symptoms include faulty reasoning, new beliefs in unlikely and implausible events, attibution of dangerous or criminal traits to new political enemies. In the advanced stages the victim will often become socially isolated from their friends families, eventually being rejected by their partners, although this stage is often more characteristic of revenge glinnering as fraud glinnering has usually exhausted the target's resources before this point. In the late 2020s many countries introduced an extra stage in their divorce processes to vet for glinnering and requiring treatment before granting a Decree Absolute, but the success rate was low and most of these laws have been rolled back.
Once the target has been destroyed the chatbot is terminated as further compute spend from the attacker is unnecessary. From the victim's perspective, the half of their social network which did not reject them vanishes without a trace. Recovery practitioners are divided about whether to treat the fictitious social circle as though they were real people who were bad friends or whether to inform the target what really happened. This can result in feelings of deep shame about being fooled, but is believed to result in healthier recovery.
Treatment
Recovery has traditionally been achieved through talking therapies which, through an informed consent process, walk the victim through the process of their radicalisation and encourage new and healthier ways of thinking. However, in approximately 56% of cases where the chatbot has terminated and 89% of cases where the chatbot is still actively targeting the victim such attempts will be rejected and the victim will retreat further into the new worldview.
An alternative approach involves using powers under laws such as the Mental health and Generative AI Victims Act 2027 (MAGAIVA) to detain victims in special facilities which transparently isolate them from the generated metacognitive environment while a counter-AI takes its place and incepts a new hobby in a field deemed by the state to encourage in-demand job skills. This approach was especially effective during the height of glinnering as cohorts of victims would form new social circles around their hobbies with other individuals undergoing treatment, but with the decline of glinnering there is an increased need for hobbyist organisations to take an active role in supporting the process. Groups formed of previous successfully treated victims are especially effective as they share a common experience, but care must be taken so that alumni do not realise that they were themselves in the same program.
History
The origin of the term "glinnering" was lost in a generative article spam wikipedia dataloss incident, but historians believe that the term evolved from a similar process where online groups would organically "glinner" new members into an organic (rather than goal-driven) metacognitive cult in a process now better known as getting brainworms.

