Should the government start controlling artificial intelligence? There are already two initiatives in Parliament aimed at doing precisely that. Critics say government controls would reverse the gains that Kenyans have made in adopting artificial intelligence technologies out of their own initiative.
Kenyans are already using AI for data analysis, graphic design and entertainment. Professional writers are using AI as an extra pair of eyes to improve their work, though AI is not yet advanced enough to generate credible articles by itself.
In 2022, Kenya was ranked fifth in Africa with regards to the government's readiness to adopt AI in public services. The ranking was published by Oxford Insights, which placed Kenya at position 90 globally. Kenya's score was the highest amongst the seven member countries of the East African Community.
Stephen Jairo of the Institute of Economic Affairs believes that Kenya's high score in East Africa is due to good technological infrastructure, such as fibre-cable connectivity and the rollout of 5G mobile networks.
Both technologies support AI. Furthermore, there is plenty of freely accessible data from various sources, such as the Kenya National Bureau of Statistics and the Kenya Open Data portal.
"The telecommunications infrastructure in Kenya has come a long way since the liberalisation of the sector in 1999 and adoption of technological changes over the years to date," Jairo says.
"Moreover, Kenya is second only to South Africa in the rollout of 5G network on the continent, albeit on a small scale to start with."
ETHICAL DEVELOPMENT
Of the two proposals in Parliament, the first is the Kenya Robotics and Artificial Intelligence Society Bill, 2023, currently with the Senate.
According to draft documents, a proposed regulatory body would "promote the responsible and ethical development and application of robotics and artificial intelligence technologies within the Republic of Kenya; and for connected purposes".
Celebrity Senator Karen Nyamu confirmed on her Instagram page on October 12 that the Senate Committee for ICT met the CEO of the Robotics Society of Kenya to discuss the proposed bill.
Incidentally, in the same month of October, Aldai MP Marianne Kitany introduced a motion in the National Assembly, calling for the formulation of a regulatory framework on AI.
She expressed concern that society was rapidly embracing AI without proper regulatory mechanisms, resulting in various negative consequences, such as disinformation and fake news.
"There is need to protect Kenyans from the potential AI-instigated harms," she said.
She cited privacy breaches, AI-powered fake technology algorithms, algorithmic discrimination, autonomous weapons and job displacement.
Other harms she noted were economic inequality, social manipulation and misinformation, financial market manipulation and privacy invasion.
Reactions to the proposed regulations on AI have been mixed. Critics fear an AI oversight body may start charging licence fees on individual users, which would make AI expensive. There have also been questions regarding the selection of board members for a proposed regulatory institution.
Supporters of government involvement insist that regulation is necessary to grow the AI sub-sector. Regulation would prevent the wrongful use of AI, such as by students who use it to write school essays in ways that teachers describe as cheating.
Kenya is not the first country to attempt to regulate AI. The European Parliament is finalising a law that may be the world's first detailed legislation on AI. If passed, the EU law would ban live monitoring and identification of people using AI systems that recognise faces (facial recognition).
Systems that use AI to assess people on the basis of behaviour, wealth, race, ethnicity and other personal characteristics would be banned in the EU. Such systems can discriminate against minorities and the poor, thus denying them jobs, bank loans, mortgages and other opportunities.
Communication systems that use AI to deliberately deceive people and vulnerable groups would be banned. The ban is designed to prevent AI from manipulating voter choices during elections. Voice-activated toys that encourage dangerous behaviour in children will also fall under this category.
DISCLOSING AI SOURCES
AI that generates text and images, such as ChatGPT, would have to comply with transparency requirements. These include disclosing that the content was generated by AI and designing AI to prevent it from generating illegal content (such as tips for hacking computers). At all times, users should be made aware whenever they are interacting with AI.
China, which has widely adopted AI, passed a law last August to regulate generative AI. According to the East Asia Forum, the Chinese law introduces new restrictions for companies providing these services to consumers regarding both the training data used and the outputs produced.
"A new provision specifying that development and innovation should be weighted equally with the security and governance of systems was also added," the Forum explains.
In the US, the senator famous for his role in reconciling Kenyan politicians after elections is part of a team drafting a law to protect the voices and faces of individuals from unfair use through AI.
Senator Chris Coons says generative AI has opened doors to exciting new artistic possibilities, but it also "presents unique challenges that make it easier than ever to use someone's voice, image or likeness without their consent".
If passed, the law would prevent the unauthorised use of individuals' images in performing video or audio recordings without the consent of the individuals.
A good example is the song "Heart on my Sleeve," which used AI-generated likenesses of the voices of pop stars Drake and The Week'nd and became very popular on YouTube, Spotify and other streaming sites within days.
The song was removed from online streaming services following complaints from the affected singers that they had not authorised that performance. Nobody knows the real identity of the people who did the AI music production.
As the old saying goes, the cat is out of the bag. It is impossible to undo the emergence of AI in Kenya, hence the call for rules on how it can be used responsibly. On the other hand, regulations should neither be too restrictive nor make it expensive to use AI.