Now, AI anchor is breaking news
Mumbai: If you were to watch news on the Chinese government-run Xinhua News Agency, you would, in all likelihood, encounter the world’s first artificial intelligence (AI)-powered news anchor.
Developed jointly by the news agency and Chinese search engine company Sogou.com, “he” is officially a member of Xinhua’s reporting team.
His voice has been modelled on that of a real anchor working for the agency, which also explains the facial expressions and actions that imitate a real person.
As with any machine learning technique, the AI anchor learns continuously from live broadcasting videos and can read texts as naturally as a professional news anchor.
He made his debut at the ongoing fifth World Internet Conference in east China’s Zhejiang Province, according to an 8 November post on xinhuanet.com.
The AI anchor can work through the day on its official website and various social media platforms, reducing news production costs and improving efficiency, according to the same report.
AI-powered virtual assistants, such as Apple’s Siri, Google Assistant, Samsung’s Bixby and Microsoft’s Cortana, are as intelligent as Xinhua’s AI anchor, if not more, but they do not have a human face.
Even IBM’s AI system, Project Debater, which engaged in the first-ever live, public debate with humans on 18 June at an event at the IBM Watson West site in San Francisco—delivering a four-minute opening statement, a four-minute rebuttal, and a two-minute summary— remains an intelligent machine without a face.
Xinhua’s anchor falls under the category of androids or humanoids—robots that look human, making them both exciting and scary. Examples of humanoids include Honda-owned Asimo and SoftBank Robotic Holdings Group-owned Pepper.
Androids or humanoids were made popular by sci-fi movies such as the epic Star Wars series, A.I. Artificial Intelligence, and Surrogates—which portrayed a futuristic world where people live within the safety of their homes while their humanoid surrogates fulfil their daily chores—and Bicentennial Man, in which a court eventually grants human status to a humanoid played by the late Robin Williams.
In a small way, this is already a reality. Hanson Robotics’ female humanoid Sophia is now a citizen of Saudi Arabia, becoming the first ever robot to have a nationality.
Activated on 14 February 2016, Sophia was also named the world’s first United Nations Innovation Champion by the UN Development Programme (UNDP) and will work with the agency to promote sustainable development and safeguard human rights and equality.
Sophia recently visited India during the Vizag Fintech Festival, hosted by the Andhra Pradesh government on 25 October.
Humanoids excite us but also raise the question whether AI-powered humanoids will become sentient any time soon.
The good news is that AI machines are nothing like the super-intelligent machines portrayed in movies. Sophia, for instance, is a “sophisticated chat-bot” (as described by the company’s website) that chooses from a large palette of template responses based on context and a limited level of understanding. However, Sophia also uses OpenCog, a sophisticated cognitive architecture created with artificial general intelligence (AGI) in mind. AGI is when an AI machine can emulate a human or even surpass it.
“The smarter these AIs and robots get, the more controversial things are likely to get, but this is also where the greatest benefit for humans and other sentient beings is going to lie,” Hanson Robotics’ chief scientist Ben Goertzel said in a June 2018 blog.
A November 2016 Stanford University-hosted report titled “Artificial Intelligence and Life in 2030” attempted to allay these fears when it said that “…unlike in the movies, there is no race of superhuman robots on the horizon or probably even possible.”
However, to avert such scenarios, given the rapid progress of technology, the US-based Defense Advanced Research Projects Agency (Darpa) is insisting on “explainable AI” as the way ahead. New machine-learning systems, according to David Gunning, programme manager at the Darpa Information Innovation Office (I2O), “will have the ability to explain their rationale, characterize their strengths and weaknesses, and convey an understanding of how they will behave in the future”.