Urgent.News

What's breaking now, across thousands of outlets.

AI

AI marketing research is moving from making content to choosing it, and Asian labs are leading

My ACM Multimedia calendar in Rio came colour-coded: must-see, recommended, if time. When I counted the must-sees, the pattern was hard to miss. Most came from Asian labs, from Singapore, Shenzhen, Hefei, Hong Kong and Taipei, plus the research arms of JD, Alibaba, Kuaishou, ByteDance, OPPO and Ant. A TechNode profile of HiDream.ai shows the […] The post AI marketing research is moving from…

AI marketing research is moving from making content to choosing it, and Asian labs are leading

From the Rio ACM Multimedia calendar, it became clear that most noteworthy presentations originated from Asian research labs. These labs were based in Singapore, Shenzhen, Hefei, Hong Kong, Taipei, and the research divisions of major companies including JD, Alibaba, Kuaishou, ByteDance, OPPO, and Ant. A TechNode profile on HiDream.ai, reveals that the commercial aspect of this trend begins with the founder Tao Mei, who previously worked at Microsoft Research Asia and served as a vice president at JD.com before establishing his own company in 2023.

The company's research initially focused on assuming a better model would sell itself, but later transitioned into workflow and agent layers. This exemplifies the evolution of marketing AI, as highlighted by the research in the article.

Written by urgent.news from e27's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at e27.co →

More in AI

GPT Image 2.6: Build a Two-Stage Sticker Pipeline with Preservation Gates

GPT Image 2.6 is most useful when you stop treating every prompt as a fresh generation. For developer-owned sticker or character workflows, a better pattern is: explore cheaply, select one reference…

  • Two-stage pipeline for stickers using Flare and Sunburst
  • Preserve character features while editing arm pose, expression, and bounce
  • Evaluate output against checklist before finalizing new candidate

More from Tuesday 17 November →