Developing AI Technology in the UK with Responsibility and Transparency

Developing AI Technology in the UK with Responsibility and Transparency

Training AI at Meta: An Overview

Meta is set to start training its AI models using public content shared by adults on Facebook and Instagram in the UK over the upcoming months. This initiative aims to create generative AI that mirrors British culture, history, and language, allowing UK-based companies and institutions to benefit from the latest advancements in technology. The goal is to build AI systems at Meta that represent the diverse global communities, with plans to expand into additional countries and languages later this year.

Engagement with Regulatory Authorities

Previously, Meta paused its AI training efforts in the UK to address various regulatory concerns. To move forward, the company has maintained positive interactions with the Information Commissioner’s Office (ICO). This collaborative relationship has proven beneficial, providing the clarity necessary for Meta to expedite the introduction of its AI services in the UK. The ICO has supported Meta’s approach to utilize ‘Legitimate Interests’ as a legal foundation for using specific first-party data to train its generative AI models.

Commitment to User Privacy

Meta emphasizes its commitment to user privacy in its AI training initiatives. The company has made it clear that:

  • Private messages exchanged between users on platforms such as Facebook and Instagram are not used for AI training.
  • Data from accounts belonging to individuals under 18 in the UK will not be used.
  • Only public information—including public posts, comments, and photos from adult users—will contribute to the improvement of generative AI models.

Notifications and User Rights

Starting next week, users aged 18 and over in the UK will receive in-app notifications on Facebook and Instagram detailing how their data may be used. These messages will include information about an objection form that users can access at any time if they wish to opt-out of having their data utilized for AI training. Importantly, Meta pledges to respect the choices of users who have already expressed objections by not reaching out to them again.

Improving Transparency

While Meta has always aimed for higher transparency compared to other companies in the tech sector, feedback from the ICO has led to enhancements in their objection process. The forms are now designed to be more straightforward and easier to locate. Meta is committed to honoring all existing and new objection forms submitted by users, reflecting their dedication to user rights in the evolving AI landscape.

Please follow and like us:

Related