President Ruto's speech during the 60th Madaraka Day celebrations in Embu was as usual notable for soaring oratory and compelling stage presence. Its climax was the exhortation that we should "deploy the best innovations and technologies to make Kenya efficient, competitive and prosperous."
The plant ecologist-cum-president's outstanding expressiveness could have served him well in another life as a stage personality. It also smashes for us in the sciences the condescending notion that we are a humourless bunch who struggle hopelessly with elocution. As Robert Mugabe once was to Zimbabwe, Kenya might soon claim the world's most eloquent president in Dr Ruto.
Anyhow, the speech was quite informative about the nation's upward trajectory in tech. "Kenya's legendary spirit of enterprise has entered the digital space," he announced, while revealing that nearly Sh30 billion has already been disbursed through the Hustler Fund to borrowers using zero bureaucracy, a device and airtime.
These are exciting milestones for our sexagenarian nation. Yet, what was most thought-provoking was Dr Ruto's hint that Kenya's "eternal conundrum" of low citizen-to-public office-holders confidence (caused by the latter's propensity for corruption) could be resolved through technology.
We wish to extrapolate on the presidential notion by proposing that our country - already famous for ground-breaking digital innovations - considers harnessing the ever-expanding capabilities of artificial intelligence (AI) in its assault on "our transactional efficiency problems" in government, to again borrow from eloquent Ruto-speak.
If anything demonstrates the positive impact tech can have on government operations and the possible gains of dabbling in AI, it is the way the e-Government initiative dramatically reduced the painful and endless queues citizens once endured at the KRA offices during renewal of driving licences, among other services.
The basic motivation of an AI government should be to detox Kenya of the everlasting governance challenges of corruption, tribalism, incompetence, document forgeries, and the like.
Computer algorithms are admittedly not always less biased than human actors. Many voices - including that of AI pioneer Geoffrey Hinton - have indeed taken to warning of its pernicious side, especially if integrated into societal infrastructure. Deceased theoretical physicist Stephen Hawking was particularly fatalistic, saying that AI "could spell the end of the human race."
Some clear dangers of AI include job losses, risky weapons automatisation, death of spontaneity and creativity, absence of human emotion and 'wisdom', and upsurge in laziness. Special interest groups might also seize the tool for manipulation of the masses for political and economic expediency.
But others are subtler, such as the replication and amplification of existing racial biases, which is seen in the way most image-processing software resort to Caucasian faces by default.
But then, for a country as ours with such a severe graft malady, no radical 'technological surgery' can be considered a worse option. After all, patriots before us who ventured out to slay this tenacious dragon with 'conventional weapons' returned home injured, discouraged and empty-handed.
It would all begin by any appropriate chatbot from the ever-expanding selection, such as GPT-4, being fed the general information about the ward (or county, or country as the case may be).
Choice of software should be informed by the maximum number of words it can take, ability to understand the logic of images, linguistic range (including Kiswahili and local languages), and general software IQ. (Can it, for instance, handle a question involving deep logic, such as, "Explain 'algebra' to a sevena-year-old"?).
This hypothetical government could be piloted at the ward level, the smallest government unit. Bot training should be deliberately wide-ranging to include content from the latest government regulations, policies, rules, official documents such as the constitution, amendments, statutes, budgets, people's opinions locally, nationally and internationally, contemporary events, ad infinitum. Since some Kenyans actually participated in the initial training of the hyped AI software ChatGPT, this is within reach.
Finally, the model - like Aladdin's Genie - should be unleashed to freely model developmental strategies for whatever jurisdiction it was designed. And, viola! Your ward or county finally has an inanimate potential MCA or governor!
As expected, our scenario-analysis for this hypothetical 'cloud-based' government is severely limited by space. Let us say, though, that if ever adopted, as most such 'wild' ideas ultimately are, county governors will for the first time ever have invisible PAs. Unlike their human 'counterparts', these will never crave a pay rise, and will possibly introduce a completely refreshing good-governance vibe in our counties.
Additionally, artificial intelligence models will obliterate the need for wananchi to "know anyone, bribe, or go through complex bureaucracy" in order to be served, besides being exponentially more accurate, fair, accessible and able to process humungous amounts of data at higher speeds than all the guys we voted in last year, combined.
All said, the scary spectre of AI one day mutating into what one philosopher calls a "one-stop oracle"- a know-it-all with the answers only God should be providing - is very, very real.
Dr Wahome teaches at Laikipia University. Nderitu is a student at Alliance High School and a member of the schools computer science club that is working on an AI model for the conservation of Kenya's wildlife
