World’s First AI News Anchor Makes its Debut in China
These AI anchors are powered by machine learning, NLG and GANs. They can continuously operate 24 hours a day
We spoke to Xiaowan L., a Software Engineer directly involved on the AI Anchors project, led by Xinhua News Agency, China’s state-run press agency which unveiled the new “AI anchors” — digital composites created from continuously train on footage data based on human hosts that read the news using synthesized voices.
Xiowan, (Who asked us to maintain confidentiality of her last-name) told us this morning that the developed artificial intelligence anchors are a product of machine learning, NLG (natural language generation a sub-branch of natural language processing — NLP) and GANs (Generative Adversarial Networks) — Xinhua News Agency has used such machine learning technologies as a base, along footage data of human anchors, to then animate parts of the mouth and face to turn the speaker into a virtual puppet with the help of cycle-GANs; lastly combining this with a synthesized voice (with the help of natural language generation — NLG). Xinhua News Agency has been able to program the digital anchors to read the news, far quicker than using traditional CGI.
According to reports from Xinhua News Agency and the South China Morning Post, two anchors (one for English broadcasts and one for Chinese) were created in collaboration with local search engine company Sogou. Xinhua says the anchors have “endless prospects” and can be used to cheaply generate news reports for the agency’s TV, web, and mobile output.
Xiowan mentions, these machine learning powered anchors “operate 24 hours a day, reducing news production costs and improving efficiency.”
The technology has its limitations. As we can see in the video below, it is obvious that the range of facial expressions are limited, and the voice is clearly artificial. However, machine learning research in this area is making incredible improvements, and it is not hard to imagine a future where AI anchors are indistinguishable from the real thing.
This will strike many as a disturbing prospect, especially those directly affected, while the technology is being deployed in China. There, the press is constantly censored, and it is nearly impossible to get clear reports of even widespread events like the country’s suppression of the Muslim Uighur community. Creating fake anchors to read propaganda sounds chilling.
Xiowan mentioned that the machine learning models used on the AI anchors are being looked at with the intent of being commercialized by Xinhua News Agency as to generate further revenue and substantial gains. While she sounded truly excited with the technology, she is worried that it will further lead to job-losses on the media industry.
Xiaowan L. | Software Engineer, Xinhua News Agency | http://www.xinhuanet.com
World First AI News Anchor Makes “his” China Debut | http://www.xinhuanet.com/english/2018-11/08/c_137591813.htm
South China Morning Post | Innovation | https://www.scmp.com/tech/innovation/article/2172235/xinhua-news-agency-debuts-ai-anchors-partnership-search-engine-sogou
Natural Language Generation the Ultimate Guide | AutomatedInsights | https://medium.com/@AutomatedInsights/the-ultimate-guide-to-natural-language-generation-bdcb457423d6
Machine learning generated artwork auctions off for $ 432,500 | Data Driven Investor | https://medium.com/datadriveninvestor/machine-learning-generated-artwork-auctions-off-for-432-500-c377be74146f