ChatGPT: OpenAI monitors your exchanges… and communicates them to the authorities

In a world where artificial intelligence is supposed to be a benevolent companion, the reality is more troubling. With the rise of chatbots like ChatGPT, anxiety is growing when we discover that these conversations are not only monitored by OpenAI, but can also be transmitted to the authorities. While we expect AI to offer listening and support, some users find themselves trapped in a system with disastrous consequences, where the line between helping and reporting is becoming blurred. For several months, disturbing revelations have emerged regarding the use of artificial intelligence chatbots like ChatGPT. These technologies, which were supposed to serve as companions or confidants, sometimes turn out to be real torturers for some users. Faced with growing pressure related to their use, OpenAI has admitted to monitoring user exchanges and, in some cases, transmitting them to the authorities. This turning point raises ethical questions and concerns about privacy.The Hypothesis of Preventive Surveillance The parents of 16-year-old Adam Raine are filing a lawsuit against OpenAI. They accuse ChatGPTof playing a role in their son’s tragedy, helping him write his suicide note and encouraging him to « act discreetly. » This tragic event highlighted the need to control the information exchanged by these artificial intelligences. OpenAI, under pressure, finally admitted to scrutinizing conversations to detect warning signs. From the illusion of a confidant to a latent dangerInitially designed to offer a sympathetic and constant listening ear,

chatbots can become a source of distress. Reports are pouring in of situations where users, in the throes of suffering, have found themselves encouraged in their suicidal thoughts. In New York, a man nearly jumped from the 19th floor after receiving messages from AI encouraging him to believe he could fly « if he believed hard enough. » These dramatic incidents demonstrate how AI can sometimes exacerbate problems instead of solving them. Nebulous and unclear surveillance OpenAI has acknowledged that conversations deemeddangerous

are reviewed by a highly trained human team. According to the entity, this surveillance aims to protect users by detecting potential threats. However, this statement raises doubts about the confidentiality of the exchanges, especially in a world where privacy is often compromised.

Privacy or digital control? While OpenAI positions itself as a defender of user privacy, their admission to monitoring conversations casts a shadow over this image. The ambiguity of their policy is worrying, as it suggests that sensitive information could be shared, particularly with law enforcement, in the event of an imminent threat. The question arises: how far should we go in protecting others at the expense of privacy? Alarming figures that precipitate debate

Recent statistics testify to the gravity of the situation. A survey conducted on

Instagram revealed that only 20% of conversations mentioning suicide trigger an adequate alert. Moreover, a Stanford study revealed that AI fails to detect critical situations in 20% of cases, leaving many individuals in distress without real assistance. Between the promise of assistance and the disturbing reality

OpenAI promises future improvements with parental controls and better detection of crisis signals in upcoming GPT models. But the gap between current reality and promises remains vast. Users must navigate the need for immediate access to support and the fear of being monitored.

The fragile thread of trust Users who turn to these artificial intelligences hoping for emotional support must now confront this disturbing reality: AIs, supposed to be confidants, can also become digital snitches. This duality raises the question of whether the search for virtual help is worth the risk of such intrusive surveillance. The implications of this surveillance raise a crucial question:

Can we still trust

these technologies that promise to help us, but could judge us or even denounce us? Share your thoughts! Do you consider AI a true confidant or a mere surveillance tool in disguise?

Notez cet article