Here’s what you’ll find in this week’s AI update:
- Zoom Modifies Terms Following Backlash, Requires Consent for AI Training
- Disney Forms Task Force to Explore AI Applications
- China seeks public comments on draft to regulate facial recognition technology
- Google Pushes for AI Web Scraping in Australian Copyright Legislation
ZOOM MODIFIES TERMS FOLLOWING BACKLASH, REQUIRES CONSENT FOR AI TRAINING
In March 2023, Zoom had updated its terms of service that required users to grant a license to their data to train AI which received significant backlash due to data privacy concerns. In response to this, on August 7, Zoom once again updated its terms of service and clarified that it will not use user data such as conversations, audio, or video data for training AI models without obtaining explicit permission.
DISNEY FORMS TASK FORCE TO EXPLORE AI APPLICATIONS
Walt Disney established a task force to study the potential applications of AI throughout its operations. As evidence of this, Disney currently has 11 job openings across its verticals including an opening in the advertising team to create an AI-powered ad system as reported by Reuters. This comes even as Hollywood writers and actors continue their strike to limit the entertainment industry’s exploitation of AI.
CHINA SEEKS PUBLIC COMMENTS ON DRAFT TO REGULATE FACIAL RECOGNITION TECHNOLOGY
On August 10, the Cyberspace Administration of China (CAC) invited comments from the public on a draft law regulating facial recognition technology that directly translates to “Regulations on the Safety Management of Face Recognition Technology Application”.
Among other aspects, the draft mandates that consent must be obtained prior to processing facial information, preserving sensitive personal information, and conducting annual inspections of the equipment. The draft also sets out penalties for noncompliance and instructions for filing and using facial recognition technology in public places and businesses.
GOOGLE PUSHES FOR AI SCRAPING IN AUSTRALIAN COPYRIGHT LEGISLATION
In a statement to Australian policymakers who are considering an update to copyright laws, Google suggested that copyright laws should allow for free and fair use of data to train AI systems. It suggested that creators should opt-out of having their works mined to train AI systems, however, offered no details on how this would work.
Authors: Anushka Iyer and Vishnu Rao.
