New Delhi, July 17: Technology companies such as Apple, NVIDIA, and Anthropic have been accused of using YouTube video transcripts from popular creators such as MKBHD and MrBeast to train AI models without their permission. This type of behavior has reportedly raised ethical concerns about the transparency and legality of the methods these AI companies use to collect training data. Using public content for AI training without the creators' explicit consent has become increasingly problematic, according to multiple reports.
Reportedly WiredApple, Nvidia, and Antropic are alleged to have used YouTube transcripts from popular YouTube creators, including MKBHD, MrBeast, Jacksepticeye, and PewDiePie, to train their AI models. The companies used YouTube video transcripts to enhance their AI models without the consent of the creators, according to a report by research firm Associates. Times NowThis use of data without consent has prompted discussion about the need for clearer regulation and guidelines in the AI industry to protect the rights of content creators. YouTube New Feature Update: Google-owned platform introduces AI-powered “look-ahead” feature for premium members. Read more.
Marques Brownlee highlights Apple's use of AI in YouTube videos
Marques Brownlee, aka MKBHD, has a YouTube channel. On July 16, Marques Brownlee posted on X that “Apple gets data for its AI from multiple companies, one of which scraped a ton of data and transcripts from YouTube videos, including mine.”
There is a dataset called YouTube Subtitles, which contains transcripts from educational and online learning channels from Khan Academy, MIT, Harvard, etc. Companies such as Apple, Nvidia, Anthropic, and Salesforce reportedly used subtitles from 173,536 YouTube videos taken from over 48,000 channels. These companies are reportedly using the dataset compiled by EleutherAI. Instagram New Feature Update: Meta-owned app introduces new stickers for Stories and Insta Reels.
EleutherAI reportedly published a research paper mentioning a dataset called Pile, a non-profit release that includes a diverse collection of material not only from YouTube but also from the European Parliament, the English version of Wikipedia, and a collection of emails from Enron employees made public during a federal investigation.
(The above story first appeared on LatestLY on Jul 17, 2024 01:18 PM IST. For more news and updates on politics, world, sports, entertainment and lifestyle, log on to our website latestly.com).
Share now
