快猫短视频

From doing backflips to public debating, robots have been grasping new tasks at an impressive rate of late. And one such task drawing increasing attention is helping children with autism to improve their social skills.

The logic goes that robots can be programmed to be consistent in their responses, removing the unpredictability from a social exchange that some autistic people struggle with. On top of this, a robot can be tailored to respond in the way judged best for each individual and programmed to incrementally increase the complexity of its interactions, making it a good social coach.

Adding credibility to this idea is a new study claiming that children with autism showed improved social skills after coaching sessions from a robot.

Brian Scassellati at Yale University and his colleagues arranged for nine children with autism between 6 and 12 years old to complete a 30-minute session with a robot every day for 30 days. The robot helped the children to play different games on a touchscreen that were designed to teach emotional and social understanding.

For example, in one game the robot would tell a story and ask the child to choose what they think different characters on the screen were feeling at a given point. Over the course of the experiment, the difficulty of the games was increased to cater for the child鈥檚 progress. So, in the case of the storytelling game, the narratives became longer and more complex.

贵耻苍听濒别蝉蝉辞苍蝉

Throughout the tasks, the robots would maintain eye contact with the child to show engagement or would turn to look at the screen when the child was doing the same. The sessions were conducted with caregivers at home.

Scassellati鈥檚 team found that after the experiment the children had improved social skills. This was measured by tests of joint attention 鈥� the ability to focus on an object by picking up on social signals from another person 鈥� and reports from their caregivers, who said the children were engaging in more social behaviour, such as eye contact and responding to communication. An improvement was still evident a month after the sessions had ended.

鈥淭he hope of this is that robots might provide a cost-effective, fun and engaging therapy,鈥� says Scassellati.

However, like many studies looking at robots and autism, the number of participants in the study is so small that it is hard to draw any definitive conclusion.

鈥淭he absence of studies with large samples makes it very difficult to know whether any of the current robotic systems 鈥榳orks鈥� for autistic children, and what it 鈥榳orks鈥� on and why,鈥� says Liz Pellicano at Macquarie University in Sydney, Australia.

Lack of control

In medicine, randomised controlled trials have long been the gold-standard for testing interventions. In the case of using robots, this would involve randomly putting children into two groups. One group would complete the robot coaching and the other would do something similar without the robot. That way any differences that show up between the groups can be attributed to each method.

鈥淲hat we want to know is whether this sort of robot-assisted interaction is at least as good as or even better than a human-led interaction 鈥� and, in particular, for which sort of children, since robots won鈥檛 necessarily appeal to every autistic child,鈥� says Pellicano.

But in this study, and many others, there was no control group, making it difficult to determine if the changes seen were simply the result of trying something new or were actually due to the robot.

Scassellati and his colleagues recognise many of these shortcomings, saying that in the future they would like to introduce randomisation and control groups to the set-up, as well as extending the length of the study.

鈥淎lmost all of the social robotics work with autistic children doesn鈥檛 really stand up to scientific test,鈥� says Pellicano.

Robots could be excellent social coaches for children with autism, but unfortunately nobody has yet done the studies to really find out.