[ad_1]
Increased Internet connectivity may not translate into increased Internet accessibility, mainly because Internet-based tools such as speech recognition are not available in most African languages - according to a collaborative analysis by the consulting firm. Caribou Digital Consulting and Digital Financial Services Laboratory.
Much has been said about the rapid spread of Internet connectivity in Africa. According to some studies, Africa will have up to 1 billion Internet connections by 2022, which should boost the digital ecosystem of the continent in general. One of the reasons for this rapid expansion of Internet access is the sharp drop in the cost of Internet connections.
The low cost of the Internet has created a more equitable environment, as people from all socio-economic backgrounds can now make the most of digital applications. However, a new analysis shows that there are still obstacles to a truly fair Internet environment.
Farnham-based consultancy and investment research firm Carabou Digital conducted a comprehensive analysis in collaboration with Digital Financial Services Lab, highlighting the language barrier that appears when most Internet applications are not developed only in a handful of languages.
Speech recognition technologies such as Alexa by Amazon and Siri by Apple are rapidly gaining popularity, so much so that 50% of all search features will be voice-based by 2020. However, like most other Internet tools neglect most African languages.
To date, there are more than 98 million Swahili speakers in Africa and more than 63 million Hausa speakers. Other languages spoken by tens of millions of speakers on the continent include Yoruba, Orkolon, Zulu, Igbo, Amharic and Berber, which are mostly absent from Internet applications.
According to the firm, this situation seems far from rectified, as most companies developing Internet-based tools continue to focus on developing their English language skills. One of the reasons for this gap in priorities is that Anglophones in Africa are the highest income earners, which is the most profitable area.
Nevertheless, there are also more practical barriers to the development of such technologies to work with other African languages, including the lack of available data to develop machine learning in more distant languages. The development of these technologies requires about 100,000 recorded speech hours, which is lacking for most African languages.
Source link
