Privacy Concerns have always remained part of the Generative AI services since its short lifetime. These concerns came into the limelight when OpenAI introduced ChatGPT and Google introduced its Bard. The learning methods and data for these tools have always remained a concern for the users and lawmakers. However, Meta has detailed its protective measures to protect User Privacy and how it collects data for the Generative AI tools that it launched recently along with 28 AI personas chatbots.
Must See: Meta launches new Ray-Ban Smart Glasses with live streaming, 12MP camera, and longer battery life
There are some privacy checks that Meta has established to safeguard user privacy in the new Generative AI tools. Following the previous check from the development language models like Llama 2 and others, Meta has further strengthened the “privacy safeguards and integrity mitigations” while building new tools and services based on AI.
From Where Meta Generative AI Tools Fetch Data?
To better train Generative AI tools, different sets of enormous data sets are required for better information and reliability of such tools. However, feeding such large systems with by eliminating privacy concerns has always remained a task. But the Beta has achieved it by using only public data from Meta services and products. It has excluded some websites where the user shares their personal information. For Eg: On Facebook, the company has processed only the data that is labeled as public, while it has not even touched the content under the private tag.
Even the data on messages are not touched while training their AI generative Tools. However, if the user asks anything from Meta AI by adding @Meta AI in their message, then only that message is processed by these tools. Otherwise, all the messages remain private. In addition, the company uses data from AI Stickers in your chat to improve these stickers.
Even, the businesses interacting with Generative AI tools such as Meta AI are part of the data sharing pacts. The company fetches the data as presided over by their new Privacy Guide.
Why does Sometimes AI Retain Information?
While using the Generative AI tools, these may retain some of your personal information to personalize the responses. It may be helpful for the users as it may help in generating more precise answers to your queries. However, there are always safeguards to protect your private information appearing in other’s responses. While talking about such issues, Meta details that “We use automated technology and people to review interactions with our AI so we can, among other things, reduce the likelihood that models’ outputs would include someone’s personal information as well as improve model performance.”
Built-in commands to Delete Private Data in Meta AI
To further equip users with better tools to delete their private information, Meta introduced some built-in commands to delete the information.
- /reset-ai – Deletes AI messages.
Further, the generative AI tools keep the account information on different Meta platforms – Facebook, Instagram, and other Apps separate.
Meta aims to protect user privacy and even has dedicated work members to abide by these privacy policies. Further, the Meta has signed the White House’s voluntary commitments for AI and is a participant in AI’s Synthetic Media Framework.
Related:
