Twitch, the popular streaming service owned by Amazon, is under fire from its community after it was revealed that user content is being used to train Amazon's artificial intelligence models by default. The setting, which allows Amazon to use data generated by creators and audiences for AI training, is enabled automatically, and users must manually disable it if they wish to opt out.
The announcement of the opt-out feature on Wednesday drew sharp criticism from users, many of whom questioned why the data collection was ever implemented without explicit consent. Twitch's chief product officer, Mike Minton, acknowledged the controversy during a livestream, admitting that an opt-in model would likely result in very few participants. "If it's opt-in, nobody would opt-in. That's the honest answer," he said.
According to Twitch's FAQs, content from channels that remain opted in could be used to train generative AI models, which are systems capable of creating new text, images, and video. Examples include chatbots like OpenAI's ChatGPT and Google's Gemini. Twitch has stated that audio data might be used to improve speech-to-text models, potentially enhancing automatic subtitles on Twitch streams and Amazon videos.
Some users have raised concerns about the implications for game developers, as the feature could also train AI on the video games being streamed. One streamer expressed frustration on social media, calling the default setting "criminal."
Twitch's head of community, Mary Kish, provided a tutorial on how to disable the feature through the Streamer Dashboard's Security and Privacy settings. However, some users reported that after toggling the option off, it appeared to revert to being enabled when they checked again.
Kish acknowledged that data collection for AI training has become standard in the industry, but she did not expect a positive reaction. "We don't expect you to be happy or excited about this. I don't expect anyone to react to this favourably," she said.
The livestream's chat was flooded with critical comments, with one viewer stating, "Nobody would opt in because nobody wants to feed AI with our creativity and content." Another suggested Twitch had an opportunity to set a better industry standard.
It remains unclear when Amazon began collecting Twitch data for AI training. Minton said he was unsure whether user data had already been used in model training. Amazon acquired Twitch for nearly $1 billion in 2014 and has invested heavily in AI development to compete with other tech giants.
Source: BBC News
មិនទាន់មានមតិយោបល់នៅឡើយទេ។ សូមធ្វើជាអ្នកដំបូងដែលចែករំលែកគំនិតរបស់អ្នក។