Recently, Discord announced five new AI features that will be coming to its platform powered by OpenAI and Midjourney, the manufacturers of the most popular AI feature, chatGPT. With this latest development to their platform, Discord promised that they would not use any of the data collected by the AI features for AI tech model training.

As part of the new features, Clyde, Discord’s existing chatbot, has now been reinforced with artificially intelligent language parsing abilities. However, there were strings attached to these updates, which made Discord readjust its initial privacy policy.

The initial privacy policy of Discord says that Clyde and other Discord products do not store the content of video or voice calls or channels. Aside from that, it also does not keep the streaming content of its users whenever they share their screens.

Discord later removed and dismissed this initial privacy policy not to collect contents of video chats or voice calls, or channels. These subtle changes were spotted and reported by Techradar on the 11th of March, with Reddit users following suit.

Discord users did not take the revocation of the promises made by the Discord platform lightly, as they immediately called them out. However, a day after they readjusted the privacy policy, Discord revised and undid those changes.

According to a statement made by a Discord spokesperson to Techradar:

“We recognize that when we recently published the tweaked language in our privacy policy, we were inadvertently confusing our users. Nothing has changed, and we’ve reintroduced the language into our Privacy Policy, along with additional clarifications.”

For further clarity on the recent changes made by Discord, Gizmodo asked about the reasons behind the changes to the privacy policy, but they were silent on the details.

Still, with the changes made to the privacy policy by the Discord platform, they promised not to keep their users in the dark regarding any information. Regarding the privacy policy, Discord will notify users of any changes regarding video chats or voice channels.

As a backup to the revision of the Discord privacy policy, Discord made some updates to its initial blog post that announced the additional AI features and integrations following the proposed changes made by TechRadar.

The blog post further clarified that,

“Clyde does not store, record, or use any voice channels or video chat data from users.” Discord added a paragraph to its blog post to re-emphasize Clyde’s “privacy-first and optional-only approach.” 

The new Discord’s AI features involve several features and tools, which include an AI-powered Cylde chatbot, which can now have conversations and answer simple questions, among others. 

There is also the AutoModAI, previously known as AutoMod, which they have now modified for large language model technology to block users’ content that contradicts the server’s rule and simultaneously adds the context of conversations. 

Another feature is the Conversation Summaries, a new AI feature for paraphrasing and restructuring many messages. These new features are like those in other Discord products that only store and use the information according to the privacy policy.

Although the Discord platform now uses OpenAI features, it is still not allowed to access Discord users’ data to develop other AI models. Discord’s chatbot, Clyde, can only access user messages when users interact directly with the AI.

One truth we all know about AI is that it constantly needs its users’ data to keep its wheels turning. These data can, however, be obtained from companies’ AI algorithms or by feeding chatbots. 

Given that large these new AI features are developed on large datasets generated by users, it is still difficult to comprehend that these OpenAI features will work without analyzing and storing data obtained from Discord users. Since Discord has given its final word on its privacy policy, users’ data should be safe and secure.