Snap vs Great Britain: Battle for Children’s Confidentiality

SNAP fell on the Radar of the British Information Affairs Directorate (ICO) due to fears that its artificial intelligence in the form of a chatbot may pose a threat to children’s confidentiality. ICO announced the issuance of a pre-warning SNAP regarding the “potential deficiency in assessing the risks of confidentiality represented by its generative chat boot ‘My ai’.

ICO actions are not related to violation. However, the notification indicates the fears of the regulator that SNAP may not take steps to ensure the correspondence of the product with the rules of data protection, which have been tightened since 2021, including the Children’s Design Code.

ICO stated that the preliminary results of their investigation indicate “an alarming misunderstanding by SNAP in sufficient identification and evaluating the risks of confidentiality for children and other users before launching ‘my ai'”. Information Commissioner, John Edwards, emphasized that organizations should take into account the risks associated with AI, along with the advantages.

Chat Bottom on the basis of AI was launched by SNAP in February, using the technology of the large language model Openai’s ChatGPT. He was presented as a virtual friend to whom you can contact for advice or send a message. Despite the declared security measures, there were reports of the incorrect behavior of the bot. For example, a chatbot proposed ways to hide the smell of alcohol to a 15-year-old user.

In response to the ICO notification, the SNAP representative said that they carefully consider the preliminary decision of the ICO and seek to protect the confidentiality of their users.

This is not the first time that the chat on the basis of AI is faced with the attention of European confidentiality regulators. Earlier, the Italian regulator issued the order to stop the processing of these local users of the Replika Chat Botu, also expressing concerns about risks for minors.

Regulators actively and publicly respond to questions related to generative chat bots based on AI, trying to ensure the compliance with the existing rules on confidentiality.

/Reports, release notes, official announcements.