Two events I attended over the last few weeks inspired this post. The first was Microsoft’s MVP Summit which they held in person for the first time in four years. The second was a Zoom call for a product called getmee which uses AI to help coach people in their communication skills.
Let us start with ‘Summit’. While the contents of the conference are under strict NDA, I do not think I am revealing too much by saying AI technologies had a strong presence. After all, Microsoft Copilot for Office has been announced, as has Copilot for Bing. In both these cases, a conversational AI bot (you may know it as ChatGPT), has been added to Microsoft applications to provide an ‘assistant’ while you work. If this sounds like the Clippit of old, while I am sure AI-purists would cringe at the comparison, it is a fair analogy. While you create an email, Copilot offers to generate the body of the email for you, adjusting the style of writing as you see fit.
I have no doubt Copilot will bring massive productivity gains to the office.
In the case of getmee, this is the brainchild of Balendran Thavarajah, a former client of mine from a little over ten years ago. Back then ‘Bala’ was a quietly-spoken, friendly development lead for a not-for-profit organisation I was working with. Today, he is the founder and CEO of getmee, a ‘personal AI coach that teaches users to speak more clearly, directly and effectively.’ Essentially, Bala had been inspired by his personal experiences arriving in Australia as a refugee with limited English. In his Zoom presentation he spoke about having sought-after skills in the market but felt his communication skills at the time hindered his ability to land roles. To ensure this did not happen to others, he created getmee.
As with Copilot, I am very excited about getmee, and products like it. I have no doubt getmee will empower people to pursue their dreams, and be more satisfied in their work.
There is an elephant in the room though with these technologies which is glossed over and is the seed for this article: products derived from ChatGPT are primarily for the English-speaking market. While it is true ChatGPT supports over 50 languages, as the source material for a language diminishes, the harder it is for an AI bot to be trained on it.
English is an excellent source of training material because it is the lingua franca of the internet and for code development. Literally over half of the material on the internet is in English i.e. there is more English content on the internet than all other languages combined.
As AI applications become more specialised, the relevant training material diminishes and, while, this is less of a problem for English due to the enormous amount of content available, it becomes a significant problem for other languages. Of course, for many countries whose primary language is not widespread, the citizens have already learned to adopt the language of their neighbors, and, what is arguably the language of business in much of the world, English. My concern is these ‘regional languages’ will become a cultural afterthought, accelerated by tools which promote excellence in productivity and communication in English.
It is a difficult problem, and certainly not the fault of the products mentioned above, which serve very real needs in English-speaking countries and countries who conduct business in English. Elsewhere though, people will be left behind. India is a great example. It is true India uses English for the language of business but only around 20% of the country speak some level of English (the percentage varies between 10-30% depending on the source.) Hindustani or ‘Hindi’ is much more widely spoken with at least 50-60% of the population speaking it to some level. In the northern half of India, this percentage is much higher and approaches 90-100% of the population. However, Hindi only makes up about 0.1% of the content on the internet.
This is why I describe AI as oscillating the digital divide. A product like getmee clearly benefits citizens in countries where English is the language of business, providing access to technological jobs which would be otherwise inaccessible, narrowing the digital divide and benefitting all. However, for a country whose citizens have limited English, they will be locked out of this new generation of productivity tools, widening the digital divide.
It would be lovely to end this article with a simple solution to address my concerns but big problems rarely have simple solutions. English language skills will continue to be the price of entry for many professional jobs around the world, and continue to be the basis for much of the world’s internet-based technologies. Where there is opportunity is in recognising this and supporting countries who are likely to be hampered the most by this trend. In the case of India, finding those rich veins of Hindi content and digitising them for use by AI will make crores of rupees, if not kharabs, for those willing to pursue and commercialise it.
That CRM Blog 