The Conversational Internet is an inspiring project developed by a team of Extreme Blue interns throughout the summer at the IBM Hursley Lab in the UK. The Royal London Society for Blind People approached IBM with the aim of creating improvements in the way that people who are blind interact with information on the Internet and the team is working towards a smart solution.
One challenge of available technology which works by reading from the top left hand corner of the page downwards, often taking 4-5 minutes to cover the whole page, is that a lot of the information that is read out is irrelevant or inapplicable to the user.
The objective is to create software that intelligently communicates information from the page to the user by imitating a conversation. This style minimises learning time for them to interact effectively with the software, as it allows the user to ask questions in natural language instead of using set commands. Users can ask questions in multiple formats, receiving results that they can act on, allowing for a smoother flow of information.
In addition to minimising the irrelevant speech output, an acceptance notification sounds whenever the software receives a command (communicating, in a clean way, that the user’s request is being processed) minimising irrelevant processing of sound by the user and maximising feedback. Software capabilities and features have also been developed with this concept in mind.
User experience testing highlighted that user preferences varied. Taking this into consideration we are developing features that enable users to edit text size (for visually impaired people) and the voice with which the software speaks . Further user interface adaption will be incorporated into the final version of our prototype.
This project is a potential area for the application and showcase of IBM’s technology. In terms of developing software capabilities, IBM’s machine learning technologies could be used to identify patterns and trends in the way web pages are structured and organised, building models that interpret a wide range of websites without manual interaction. IBM’s Content Analytics also play a huge role in the potential for more intelligent categorisation of information and therefore feedback to the user. If developed effectively, the potential for this product to empower its user is immense.
The team have been able to validate the need for this project and create a proof of concept that qualifies our approach. We have confirmed that these technologies can make a difference in the way that blind people experience the Internet. This project has been fast paced and exciting; 7 weeks of development, 3 phases of user experience testing and collaboration with departments in Hursley, including weekly feedback presentations, have all contributed to our dynamic approach. With few weeks of development left and a vast pool of possible paths, your feedback, suggestions and questions would add huge value to our project. You can contact us at : email@example.com