The THiNK founder Brian Omwenga moderated the panel session on Artificial Intelligence(AI); the economy of tomorrow.
This was the second time the THiNK founder had the honour of moderating this session, with immense gratitude to the ICTA team’s invitation. Brian was part of the lead consulting team that put together the National Digital Master Plan 2022-2032; he drafted the plans for the emerging technologies pillar which covered Artificial Intelligence. This year’s summit offered a good opportunity to review and evaluate progress around the masterplan, and in this particular case AI in Kenya. The panel also took a dive into evaluating the AI tech ecosystem and its challenges within the Kenyan context.
AI is Present, Alive and Vibrant in Kenya – The Panel
The panel discussion took place on the second day of the summit, in the early afternoon. The discussion was opened up with a key note from Dr. Joseph Sevilla of Strathmore iLab. He centered his discussion around Natural Language Generation (NLG) models, and in particular ChatGPT which has been all the rave in the last few months. He cracked up the audience when he discussed the attempts of one of his friends to crack Chat GPT and the excitement the friend would get when sending ChatGPT into so-called “hallucinations”. As a veteran in the technology space locally, Dr. Sevilla, also reminded us that developments in computing, which at first may seem scary, are often misunderstood. He encouraged us not to be driven by fear, which may result in our lack of participation in an advancement that may also present immense benefits. He jokingly added that ChatGPT had already started creating new jobs such as a ChatGPT prompter. The key note was followed by a Case Study presentation by the team at UiPath. UiPath is a leading company deploying AI solutions, and in particular, the team took us through Robotic Process Automation (RPA) solutions that they have been able to deploy in various governments and government related agencies with extremely good results. What then followed was the panel discussion. Japhet Murray set things rolling by explaining that the proliferation of AI has been caused by the recent reduction in compute cost, making it a lot more affordable. He explained this from the perspective of his company which provides infrastructure solutions, which has seen a drop in compute cost over the years. His counterpart Joe Kanyua from Oracle re-iterated this, adding that the added availability of data also recently made it possible to train these AI models, hence the added availability of AI. He stated that the real use case of AI is to augment human intelligence which can have good business outcomes. He gave several other use cases that were also picked up and explained by Prajwal from UI Path. In the discussion around the challenges of setting AI enterprises in Kenya, Jessica Colaco gave her journey setting up her AI startup Brave. This was followed by a shout out by Brian to the NLP team at THiNK, and further discussion by Prof. Audrey Mbogho about the challenges experienced in training students to become AI professionals within the sector. Florence Munene from Down to Earth technologies brought us down to earth and took us close to the edge by informing us about the uncertainties that AI also presents. A most profound submission was made by Martin Mbaya from Strathmore, who, in his attempt to explain ethics and the importance of people in the AI discussion summarised his considerations in the acronym M.B.O.N.A. We borrow from the published post of Deshni Govender below:
Mbona is a swahili word for ‘Why?’ and as an acronym, it aptly summarized the panel discussion on how AI will impact the future of work.
M: Me… that AI is a tool for me, you or institutions either as users or as the developers of AI and the ethical issues therein. That is, AI is a tool that augments functions of humans; but it is probable that humans will need to reskill and take up new functions and roles.
B: Bias… AI is biased; it might not only respond according to the quality and context of the data it has been trained with; but increasingly it also mirrors the values of its creators.
O: Ownership… The concept of ‘digital bandits’ raised broadly in the conference; bordering on how know-how is appropriated and who is the ‘consumer’ versus who is the ‘producer’ of AI solutions within East Africa and broader African continent. This also touches on ‘how’ the topic is taught to students and whether they see it as an opportunity for the future or they would just want good grades.
N: Natural… Artificial Intelligence ultimately depends on natural intelligence. Natural intelligence is powered by the human brain (neurons and synapses) and the brain is complex. Emotion, empathy, etc… all natural to human beings are very very difficult to build into machines. The machine is therefore a tool that augments humans (and vice versa) … or is it?
A: Artificial… Artificial Intelligence is just that, ‘artificial’. Artificial intelligence is powered by machines… central processing units (CPUs) and graphics processing units (GPUs) and relies heavily on faster (and more affordable) speeds of compute that are enable AI to go into the hands of even more users. The final contributions were made by Mike Odhiambo the Director of Shared Services at ICTA. He highlighted the role of regulation, and particularly the focus of caution in use cases of high risk in the application of AI. He ended by underlining the importance of localization and training of models, to avoid the instances of bias. It was a highly diversified panel, that indeed gave a masterclass that touched on infrastructure, finance, language, governance, regulation, market, enterprise, tools, applications, data and most importantly people. We look forward to checking on more progress in the coming years.
Show Case
THiNK had the good fortune of describing their multilingual chatbot project with the Business Registration Service in the 2022 Connected Summit. This year, the team at THiNK showcased the multilingual chatbot they are developing for the Office of the Data Protection Commissioner. We believe that this chatbot will go a long way to assist the Commissioner fulfill her mandate of increasing data subject awareness and civic education. But most of all, it is our hope that this chatbot not only serves, the Kenyan public, but it also inspires the Kenyan AI community of practice to keep applying themselves to build more and more localized solutions around AI. Video link of panel: https://www.youtube.com/live/Y7QoBq4tPwY?feature=share&t=18965
