[ad_1]
The Co-Founder of the Responsible AI Network (RAIN) Africa, Jerry John Kponyo is calling on relevant stakeholders to incorporate data of all groups as more Artificial Intelligence (AI) systems are being developed.
Analysts have in recent times stressed the importance of getting used to various AI systems being developed in order not to be left behind.
These systems however need to be fed with data from all areas for them to give a fair analysis with respect to issues.
The Co-Founder of the Responsible AI Network (RAIN) Africa, Jerry John Kponyo is thus advocating for the need to ensure a responsible AI system where data from Africa for instance would be well incorporated.
“Where there are biases-when we talk of biases, it is a false balance. For instance, if you are collecting data as far as the human race is concerned and your data is so full of people who are of a certain origin. Let’s say your data is collected with people from America, Asia or from Europe and the data does not capture people from Africa, when you feed that to the machine, it gives it an impression as though the only people in the world are the people outside Africa. So, when solutions are deployed, you are likely to get erroneous results. All that we are saying is that, where there are biases as far as the data that you are training your modules is concerned, those biases need to be corrected so that the machine is not fooled into thinking that what it has is all the information that is available,” Prof. Jerry John Kponyo stated.
He wants AI solutions being developed to be evaluated according to established standards which would be responsible, trustworthy and should not be dumped on people as the absolute truth when a section of people were not considered while developing such a solution.
“The key considerations for a responsible AI are fairness, accountability, equity, data and privacy, transparency, the social, environmental and safety dimensions as far as AI is concerned. These are the building blocks as far as responsible AI is concerned. When we say an AI solution is responsible, what it means is that, it has been subjected to these principles and has been found worthy of consumption,” Prof. Jerry John Kponyo added.
He made this statement at a conference at the Kwame Nkrumah University of Science and Technology which brought various stakeholders together on ensuring responsible Artificial Intelligence (AI).
RAIN
Responsible AI Network (RAIN) Africa, is a collaborative effort between KNUST and TUM Germany.
KNUST has thus come up with a Responsible Artificial Intelligence Lab (RAIL) inside the postgraduate building of the College of Engineering.
This is where the University enlightens students and researchers more on issues related to Artificial Intelligence.
[ad_2]
Source link