Apr 09, 2011
Voice based web browsing is the new revolution in empowering indigenous communities through the web. Stephane Boyera of World Wide Web Foundation talks to OneWorld about harnessing this new technology by using the rich mobile user base in the developing world.
OneWorld: You have been a part of World Wide Web founding team. Could you please expand the role of The World Wide Web Foundation that has played a significant role to make web access free and open for all?
Stéphane Boyera: The Web Foundation was officially launched in 2008 by Tim Berners-Lee, the founder of World Wide Activity. Tim invented the web in 1989 and as it was growing he decided that there should be a place where organisations and people work together to extend the web and its capacity. He founded the World Wide Web Consortium (W3C) that is an international organisation for web technologies. W3C has been critical to ensure the growth of the web as we know today. It is based on free and open standards and can be used by all.
The web was largely expanding in the western countries and around 2005-06 Tim wanted to study how the benefits of the web can be extended to the developing countries. We started with the first interest group within W3C that is called the Mobile Web for Social Development. We soon found out that probably W3C is not the best structure for developing nations. For this reason Tim decided to create World Wide Web Foundation.
It has three major components: a) To support the mission of a free and open web and to support the standard work that he has done in W3C; b) To support research on the web and how the web as a complex system is working. That is also the kind of work the Web Science Trust is doing. This is another organisation led by the team. And c) the Web in Society that is extending the benefit of the web in developing countries. This is the programme I am leading.
OW: Your new project on voice browsing, Voice-Browsing Acceptance and Trust (VBAT), is promising in giving web access to those who have little or nil technology skills as well as the visually impaired. Could you please throw some light on the VBAT project? What steps have you taken to ensure confidence of the users in accepting information delivered through the voice-browsing technique?
SB: This project is very important for us. Mostly we learned through Web in Society that the web was actually not empowering the developing countries for two major reasons. 1) The question of accessibility: How to access content on the website? This is related to the device, connectivity, language, literacy level, impairments, disabilities, etc. 2) the question of relevant and useful information for those living in developing countries. Even the web has billions of resources but very few are coming from Africa, Asia, South Asia, Latin America, etc.
In order to address these two dimensions, we have four initiatives in the foundation: First, on Mobile Entrepreneurship where we are trying to create mobile services that create jobs for people. And at the same time create new innovative services for their community.
The second initiative is called Open Government Data which is trying to help the government to be more transparent by exposing what they are doing for the citizens. The other way is to provide all the data so that all innovative services can use the data.
The third one is Voice Browsing. We believe that voice as a channel to access data and browse has not fully been explored and which can actually serve a lot of barriers, that is the barrier of connectivity; you don’t need to have internet connectivity. You just need to have a phone.
Today the world has 5 billion mobile phone subscriptions in the world compared to only 2 billion of people using the web. It’s a big gap that we can bridge with voice browsing. It is a technology that will deliver content in all languages of the world. It is accessible for people who are illiterate, for people who have visual disabilities. About 90% of the visually impaired people live in developing countries and the current existing technologies to assist them are very expensive.
The fourth initiative is called the Web for Agriculture. In countries like Africa agriculture provides 70-80% of employment. In these areas people are also extremely poor. We believe that availability of important information can actually strongly improve the lives of these people. That’s why we chose this domain to extend the benefit of the web to those who are most in need.
The VBAT project is at the intersection of voice browsing initiative and agriculture. This project explores different human dimensions of technology – how can people talk to a machine, how can they get information that they will trust and use in their fields. What we are going to do in this project is to include all these human factors. The second step is to integrate more of voice technology in the application so that all the information that is stored in the database is easily and directly accessible by the farmers without any human intervention. That’s our ultimate goal.
OW: What is the Text to Speech (TTS) technology involved here?
SB: TTS is one of the dimensions critical in this work. In an HTML page, you need a screen for display and a keyboard to provide machine user interface. Voice application is exactly the same. You use your ear to listen to the content and use your mouth to generate the prompt. The computer has to do the same on the other side. Either it does it by managing an audio file which is not very dynamic. So we use different modules. One of them is Text To Speech which is a way to generate audio prompt from the text.
This allows people to use the information which exists in textual form and move it to the voice world. In the other direction is speech recognition. How can the computer recognise what the farmer is asking for and transfer the voice prompt into a textual command to facilitate smooth web browsing.
A huge amount of work has been done in English, Spanish and French. Unfortunately this work hasn’t been done in languages spoken in developing countries, for example, there are millions of people speaking Hindi but hardly any seminal work in Text To Speech and Speech Recognition in the Hindi language. The quality of generated audio has also not been improved.
We found out in other studies that people don’t trust a non-human audio. They rely more on a more human-like voice. That is the project we are currently working on – to find and test different types of TTS in Hindi to find one or more of what could be acceptable and trustable by farmers.
OW: What are your expectations from the project in India?
SB: Our expectation is to demonstrate in the end that farmers can interact with computers without a human intervention; that it is possible to dynamically generate audio acceptable and trustable by farmers and to generate content through TTS in Hindi that is trustable, understandable and acceptable by the farmers.
If we are able to demonstrate though qualitative and quantitative statistical analysis, then we will be able to take a big step towards further integrating voice technologies in the environment of farmers.
OW: Any comments on your partnership with OneWorld in implementation of the project?
SB: It is a great opportunity for us to partner with OneWorld. What we have at the Web Foundation is very good knowledge and technology. We are also working with the specialists in languages, in voice technologies. We work in the University of South Africa, Northwest University. We have this world map enabling voice browsing of the web. What we are missing is a test bed environment, to work with people in the field who are already supporting farmers.
The work that OneWorld is doing with the LifeLines project is impressive. There is also an eco-system which is working. Through our partnership, I feel we can make it more scalable and sustainable so that it can benefit a larger population. For us, this is a really perfect partnership because it offers a perfect case to test our technology on one side and for OneWorld, on the other, to have more impact on the farmers. I see both partners winning a lot in this partnership and I am happy to be here to launch this meeting.
OW: What are your thoughts on the mobile impact in India? Do you think poverty reduction can be linked to the growth of affordable handsets?
SB: It has largely been demonstrated that the penetration rate of mobiles have a direct impact on the GDP. Because people are able to save on transport and interact better in the value chain.
But according to me, we have exploited only 1% of the potential. Compared to the number of websites that are launched every day, the number of mobile services reaching out into the rural areas are still very low in the developing world. It is a huge opportunity for the growth of more services – to access new customers, new markets. For me, the next revolution is to move from person-to-person communication to releasing information in the society and this is where we see the major growth potential for people living below $1 a day.
Globally, what can really affect people is information. Also we work with other traditional methods of communications – in Africa and also in India community radio is a strong media to disseminate information to people in rural areas. So for us, connecting mobile, web and radio is also very powerful. It’s important that the mobile is not the focus; the focus should be to provide information to people to improve their lives. And to exploit different channels possible so that people have access to this information.
A founding member of the World Wide Web Foundation, Stéphane Boyera is currently Program Manager of Web in Society, a new Foundation programme targeted at leveraging the power of web technologies to empower people in the developing world. Stéphane is also on different expert panels on the topic of mobile and social development, including the World Bank Expert Panel on Mobile for Rural Development, the World Bank Expert Panel on mobile Health, and Vodafone Socio-economic Impact of Mobile (SIM) panel.