AI Special with Joris Knetsch of Prose on Pixels: OpenAI’s Sora, DeepSeek, and China’s AI Labelling Law
by Podcast
on 10th Oct 2025 inIn this MadTech Podcast AI special, ExchangeWire editor Aimee Newell Tarín and COO Lindsay Rowntree are joined by Joris Knetsch, EVP at Prose on Pixels, APAC.
They discuss the opt-out model in OpenAI's Sora video generation platform, US concerns about China's DeepSeek, and regulations in China's AI landscape. The conversation examines content rights, notions of privacy and accuracy in Asia and the West, and authentic versus synthetic content.
This week's stories:
OpenAI to boost content owners' control for Sora AI video app, plans monetisation (Reuters)
OpenAI will be introducing controls allowing holders of content rights to dictate how their characters are used in Sora, its AI video generation tool. Copyright owners, such as television and movie studios, will be able to block the use of their characters. The company also intends to share some of the revenue generated with rights holders who want their characters generated by users.
AI standards institute sounds alarm over DeepSeek (Axios)
A report from the US government has sounded alarm bells over China's DeepSeek models, claiming they pose risks to national security. The evaluation compares DeepSeek models against three OpenAI models and one Anthropic model.
According to the report, DeepSeek's most secure model was, on average, 12 times more likely than the US models to "follow malicious instructions designed to derail them from user tasks”. It also said that DeepSeek models gave "4 times as many inaccurate and misleading" responses on a dataset of "politically sensitive questions."
China releases 'AI Plus' plan, rolls out AI labelling law (IAPP)
China issued draft AI ethics rules for public consultation. The rules apply to all AI research and development and services in China that could affect health and safety, reputation, public order, and sustainability. Projects which fall under these must undergo an ethics review.
China also rolled out mandatory AI labelling rules obligating AI-generated content service providers to clearly mark AI-generated content. The responsibility to comply lies with internet platforms.
Follow ExchangeWire