The algorithm which is the Artificial Intelligence (AI) that powers social media is a machine that needs to be trained and controlled, in order to avoid it spiraling out of control.
“Due to the happenings on social media, industry practitioners are pushing for “explainability in the AI”, also referred to as “black box” within the AI algorithm”, the Director-General of the National Information Technology Development Agency (NITDA), Kashifu Inuwa, said.
Inuwa who spoke while delivering a Keynote Address on “Artificial Intelligence: Effective Communication for National Security” at the 2023 Security and Spokespersons Awards and Conference which in Abuja, said reiterated the need to explore and use AI in bolstering productivity in critical sectors of the Nation’s economy.
“Imagine having a system that makes decisions for us and does not understand the way it came about making that decision, now, that is indeed a red flag, having AI go out of control, so, we need to get AI to be always for humanity, because it is the most existential threat to humanity today”, stressed.
“We are seeing how AI is becoming biased, gender biased and even exhibiting racist behaviour, partly because of the data used to develop it. So, if you develop an AI system and train it in another country and decide to take it elsewhere, it will not work, hence, the need for global data for every one use to train it”, Inuwa noted.
He said that AI is beyond a tool, as it is a new way of building a relationship between humans and machines, saying, “In building that relationship, we need to look beyond just having a tool we use for our daily activities and start thinking in the direction of developing a system which can help meet certain human needs.”
Inuwa said, “countries around the world are making efforts toward deciphering the functionalities of artificial intelligence and coming up with policies and putting standards in place to ensure they are AI ready countries, and so Nigeria cannot take the back seat”.
“In Nigeria, we are looking at it from three perspectives; having AI ready principles, data and AI ready security. This invariably means that as a matter of principle, we should have redlines that AI must not cross, to achieve this, last year we started our journey by crowdsourcing on how to build National AI Policy and we have the first draft ready”, Inuwa stated.
The NITDA Boss emphasised the need to speed up efforts to connect the unconnected in order for them to be digitally feasible ahead of the AI adoption and deployment in the country, added that if they are infeasible, the AI will not consider them while making decisions.