Seven years after our research into artificial newsreaders, China has given them bodies. But can we engineer the authority and emotion that make news believable?
In November 2026, China's state news agency Xinhua unveiled two humanoid robot newsreaders at the World Media Summit in Beijing. Named Xinyan and Xinyu, they have human proportions, lifelike faces, carefully calibrated expressions and the ability to deliver news on camera. Their appearance was overseen by Chinese film director Zhang Yimou, whose involvement suggests that the challenge is not simply one of robotics, but of character, performance and believability.
The news took me straight back to some research I conducted seven years ago.
In 2019, I authored The New Dimensions of Media, an independent report commissioned by Sky exploring five trends that might shape the future of media. One of those territories was Trusted Sources, within which we examined something I called Engineered Authenticity.
At the time, Xinhua had recently introduced an AI newsreader, a computer-generated presenter designed to deliver the news around the clock. Unlike the humanoids unveiled this month, it existed on a screen, simulating the voice, facial movements and gestures of a human broadcaster.
We showed this development to consumers as part of our research. Their responses were revealing.
One participant, Jodie, from Generation Z, put it rather well:
"It feels contradictory, something not quite real is broadcasting news which is very real."
It was an observation about much more than the technology.
News is not just information
Our research suggested that audiences were uncomfortable with the idea of an artificial newsreader, particularly when it came to serious or emotionally charged events.
A natural disaster. A tragic accident. A national emergency. These are not simply pieces of information to be communicated. They require some acknowledgement of human experience, of the gravity of an event, of the people involved and those affected.
We concluded that delivering the news was about more than delivering facts. It was also about communicating emotion, and that the authority and authenticity of the person delivering the news mattered as much as the words themselves.
Indeed, our research suggested that AI news anchoring might not be the future.
Seven years later, the technology has not disappeared. But it has acquired a body.
But does that mean our original respondents were wrong? Or have we been looking at the problem from the wrong direction?
From headline to footnote
There is another change that has taken place over those seven years, and perhaps over the previous twenty, that might be just as significant as the development of the technology.
The experience of receiving the news has changed.
I can remember when certain news stories seemed to stop the day. An announcement would interrupt normal programming. A presenter would appear, sometimes visibly shaken, to communicate something of national or international importance.
There was a collective quality to the experience. We heard the news together, or at least through a relatively small number of shared channels. The presence and demeanour of the newsreader helped convey its significance.
Of course, not every story was treated this way, and television news was never without its own conventions of theatricality. But there was a recognisable ritual of public announcement.
Today, much of the news arrives differently (or, indifferently).
An act of violence that might once have shocked a community can appear as one item among many in a continuous succession of reports. A stabbing, a murder, another attack, another tragedy. Each is devastating, but in the flow of news, one story quickly gives way to another.
Sometimes the story does not even warrant a presenter. It travels across the bottom of the television screen as a chyron, or appears as a notification on a phone. The information is delivered while we are occupied with something else.
What might once have been a headline becomes almost a footnote.
I am not suggesting that such events matter less, or that people have become incapable of compassion. But the conditions under which we encounter the news have altered. The emotional significance of an event and the amount of attention it receives are no longer necessarily aligned.
The Reuters Institute's Digital News Report 2026 describes a world in which social and video platforms have overtaken news organisations' own websites and apps as sources of online news. It also reports that 42 per cent of respondents across its markets sometimes or often avoid the news.
News has become simultaneously more available and, for many, more exhausting.
Perhaps the artificial newsreader is arriving at precisely the moment when the ceremonial function of the newsreader is becoming less central to everyday news consumption.
If the news is just another item in the feed, perhaps the synthetic presenter does not need to persuade us that it feels anything at all.
It simply needs to keep reading.
The manufacture of feeling
Yet the Xinhua robots appear to be designed for something more ambitious than the efficient recitation of information.
Their developers have incorporated bionic skin, expressive movement and an AI model intended to support emotional expression. The involvement of a celebrated film director is especially interesting. These are machines whose appearance and performance have been deliberately constructed to evoke the qualities of human presenters.
In other words, the technology is attempting to address the very discomfort identified in our 2019 research.
But there is an important difference between reproducing an emotional expression and experiencing the emotion itself.
A robot can be programmed to lower its voice when reporting a death. It can pause before delivering difficult news. It can arrange its facial features into an expression associated with sorrow. It may one day perform these gestures so convincingly that an audience struggles to tell the difference.
But does that resolve the problem of authenticity?
Human newsreaders are themselves performers in a professional sense. They are trained to modulate their voices, control their expressions and communicate appropriate seriousness. They do not need to experience personal grief for every tragedy they report. In fact, professional restraint may be essential to their role.
The distinction, then, cannot simply be between genuine emotion and performed emotion.
It is also between a person exercising judgement and a system producing the appearance of judgement.
A human presenter can be challenged, questioned, held accountable and expected to respond to circumstances beyond the script. The emotional register of their delivery is part of a larger human responsibility.
With an artificial presenter, that register becomes something designed, specified and controlled by others.
The performance of compassion is separated from the capacity to care.
When the skin becomes real
There is another development worth considering.
Researchers at the University of Tokyo have demonstrated robotic faces covered in living skin tissue, grown from human cells. In 2024, they showed how the tissue could be attached to a robotic structure and moved into an expression resembling a smile.
This remains experimental research, and it should not be confused with the bionic skin used in Xinhua's newsreaders. Nevertheless, it points towards a future in which the boundary between artificial appearance and biological material becomes increasingly complicated.
A robot might eventually possess skin made from living cells, equipped with increasingly sophisticated sensory capabilities.
The surface of the machine could become biologically real.
But living skin does not, by itself, establish the existence of emotional experience. A face can move into a smile without there being anyone, or anything, experiencing happiness behind it.
We may therefore be moving towards a curious inversion.
The more materially human the machine becomes, the more difficult it could be to establish what, if anything, is actually being felt.
And perhaps audiences will stop asking.
The news we cannot simply scroll past
There will still be moments when news becomes something more than information.
The death of a national figure. A catastrophe. A moment of collective uncertainty. An event that changes how a country understands itself.
At these moments, the person delivering the news is not merely reporting that something has happened. They are participating in a shared act of recognition.
It is difficult to imagine that a machine, however convincing its appearance, could fulfil that role in quite the same way.
But I may be wrong about what future audiences will accept.
Perhaps people will develop emotional attachments to artificial presenters. Perhaps the familiarity of a synthetic face will become reassuring. Perhaps entire generations will come to regard human newsreaders as unnecessarily unpredictable, subjective or imperfect.
And perhaps what changes most profoundly will not be the machine, but the expectations of the audience.
That is the possibility I find most interesting.
Seven years ago, our research identified a tension between the reality of the news and the unreality of the person delivering it.
Today, the artificial presenter is becoming more convincing, while the news itself is increasingly unconvincing when encountered, and devoid of the shared rituals that once helped establish its significance.
We have become better at manufacturing the appearance of humanity at the same time as we may be losing some of the collective experiences through which humanity recognises itself.
In 2019, a young woman told us that it felt contradictory for something not quite real to broadcast news that was very real.
Seven years later, the newsreaders look more real than ever.
Perhaps the question is no longer whether we believe in the person delivering the news, but whether we still experience the news as something that is authentically news, or even shared by all.
References
Tracey Follows, The New Dimensions of Media: Five Trends for the Future, Futuremade, commissioned by Sky, June 2019. Particularly pp. 24–26, Engineered Authenticity and Audience Protection.
Xinhua News Agency, 5 November 2026, announcement of Xinyan and Xinyu at the 7th World Media Summit.
Reuters Institute for the Study of Journalism, Digital News Report 2026.
Reuters Institute, Generative AI and News Report 2026.
University of Tokyo, Robots Face the Future, 26 June 2024.