Andreas Bethke, Managing Director of DBSV, on the potential that AI offers people with visual impairments.
“The scope for social innovations with AI is far from exhausted!”
Listen to the interview with Mr Bethke (in German):
Interview with Mr Bethke as text version:
Policy Lab: The German Federation of the Blind and Partially Sighted (DBSV) is supporting the Civic Innovation Platform as a cooperation partner. Why is the project of interest to your association? What benefits do you expect from the Civic Innovation Platform?
Bethke: The platform is of interest to us because it sets an example for collaboration and sets the benefits for society at its core. This gives us the opportunity to contribute to our own agenda, namely designing innovations to be accessible for disabled people. As well as this, we are familiar with the daily routines of disabled people and are able to contribute solutions and make suggestions ourselves. We think that the potential for social innovations with AI and digitalisation is far from exhausted.
Policy Lab: What do good AI applications need to offer in order to make sure they can be used by everyone?
Bethke: First of all, we need to have clear rules and regulations. The coronavirus app, for example – which, admittedly, is not AI, although it is a digital application – is accessible. Why is this the reason? Quite simply because this was expressly demanded in the tendering process. So, we have to pay clear attention to accessibility and usability when we define the underlying conditions. Then, we need to get more people qualified so that those who must ultimately implement the applications also have the necessary background knowledge. And we need to reach a higher level of awareness in order to perceivesociety in a more inclusive and divers way. By working towards this aim, we can ultimately also fulfil the promise harboured by the word “civic”, namely cooperation between different stakeholders across society. And when we do all that, we will be able to create good innovations.
"We need to reach a higher level of awareness in order to perceive society in a more inclusive and divers way. By working towards this aim, we can ultimately also fulfil the promise harboured by the word 'civic', namely cooperation between different stakeholders across society."
Policy Lab: How do you assess the general potential offered by AI for supporting people with visual impairments; blind people? Will technological innovations have a revolutionary nature similar to that of Braille, for example?
Bethke: I think they already have. One good example is the screen readers that interpret the content of displays to help blind and visually impaired people to work with computer programs that were originally only designed to be used by sighted persons. The software, which is based on learning processes, has already progressed considerably. If it weren’t for screen readers and OCR, I wouldn’t be able to do my job at all. Looking forward, voice control and autonomous driving may also have potential for advances on a scale comparable to that of Braille.
Policy Lab: An app that uses a pair of glasses to turn visual information into acoustic signals is now also available and is designed to help visually impaired people perceive their surroundings. What do you think of that?
Bethke: For years now, there have been developments that use environmental recognition and help to improve the mobility of visually impaired people. I personally don’t use them, although I do know that many people are already finding such additional tools very helpful.
Policy Lab: Why do you not use these?
Bethke: As I was born blind, I learned as a child to use my hearing and to find my way around with a cane. In other words, I’m already using my ears to perceive my surroundings. If I additionally had to acoustically process information provided by glasses, I think I would have difficulty finding the right balance. You quickly find yourself bombarded by different noises. I tend to find information that helps me to navigate more useful when, for example, someone tells me that I’m just passing a bus stop for the bus line I want or at traffic lights when I need to cross a road. The benefits of orientation aids depend heavily on the age at which you lose your sight or become visually impaired. Are you even able to train your other senses or do you depend on technical assistance? There are very diverse needs. Your own personal history and lifestyle also play a role of course.
Policy Lab: The insides of buildings, for example, pose a particular challenge for navigation systems for the visually impaired due to the absence of public data.
Bethke: Yes, that’s right. This aspect will, I think, grow in importance, as well as the orientation towards buildings. Returning to the example of autonomous driving, the question arises as to how the “final mile” is to be bridged. In other words, how do I get to the building I want to go to if I’m no longer able to ask the bus driver? There is a very strong need for the development of good AI solutions and this need will become much bigger with time.
Policy Lab: To what extent do you think AI applications will help the visually impaired become more integrated in the employment market?
Bethke: My hope is that we will be able to make progress in the automatic conversion of visual materials into accessible documents. In other words, a way in which structures such as headlines can be automatically recognised and their content interpreted as a basis for creating documents. We must also find a way of making the interface of websites more intuitive to use. Currently, people who cannot see must click their way from link to link or from heading to heading. If there were AI solutions for shortening these paths or for providing an interpretation of websites’ content, this would improve the efficiency of work greatly. Automated descriptions of images would also help.
Policy Lab: If you could have any AI application you wanted, what would it be?
Bethke: It would fascinate me if in the area of virtual realities there were haptic applications that allowed me to touch or feel things that can normally only be seen. I think that would truly be a revolution! Applications based on voice control of the type we are already familiar with from Star Trek are perhaps more realistic and closer to realisation. What I mean, for example, is a situation in which I could simply tell the computer that I want a new pair of trousers or ask what is available on the market, what the current fashion is, what it would recommend, and the computer would then provide me with a few offers. In other words, the possibility of experiencing the digital world on a purely voice-controlled basis. That would be wonderful!
Policy Lab: Mr Bethke, many thanks for talking to us!