In this blog post I would like share my experience of attending the IEEE CertifAIEd™ Assessor Training.
First of all if you are interested in registering for the training head over to this link. The training is held only every 3-4 months with limited seats, so register while you can.
Why become an IEEE CertifAIEd™ Assessor?
I'm a freelance AI Consultant working on cutting edge AI and machine learning problems and it worries me that the power of these technologies are not being assessed for harm before they are applied to use cases that impact human lives and society in general.
Hence I was looking for a framework that could help me assess AI related use cases and evaluate them from the perspective of ethics. The programs I came across were:-
Singapore Computer Society's AI Ethics & Governance Certification (Professionals).
Certified Ethical AI Professional (CEAIP) - which does not seem to exist anymore.
IEEE CertifAIEd Assessor.
Based on my evaluation the IEEE CertifAIEd™ Assessor was the most comprehensive and well designed training program among these three. This in-depth training is aimed at individuals who would like to learn how to perform assessment or certification against IEEE CertifAIEd AI ethics criteria and is a requirement to become an IEEE CertifAIEd Authorized Assessor
The program is backed by IEEE Standards Association (IEEE SA) and has been developed based on the ontological specifications drafted by industries leading experts on AI ethics.
Read more about the program here.
How is the program structured?
The training lasts a full THREE days with the exam scheduled on the FOURTH day for approx FOUR hours.
The training is quite intense and expert certified trainers cover all the ontological specifications in detail while simultaneously encouraging engagement from the participants. There are plenty of exercises which one has to solve in groups and these exercises are really helpful in preparing you for the certification exam which is organised on the final and fourth day.
The schedule and program might be subject to change but here is how my training was organised.
Day 1 - Introduction and Ethical profiling with group exercise to solve a case study for Ethical profiling.
Day 2 - Review of criterion suite for evaluating Accountability and Transparency, and group exercises on evaluating two use cases on these two criterion suites.
Day 3 - Review of criterion suite for evaluating Algorithmic Bias and Privacy, and group exercises on evaluating two use cases on these two criterion suites.
Day 4 - Certification exam which encompasses the evaluation of a use case based on ethical profiling and the four criterion suites that were discussed during the training.
The training I attended was extremely well organised and the trainers did their best to keep us engaged for three full days of lectures and group exercises.
What happens after the certification exam?
A month or two after the candidate clears the exam, the candidate is invited to complete an application, or Case for AI Ethics Competence. This online form consists of eight pages describing the competence goals and inhibitors for the role. Candidates will be requested to complete the required fields by providing or describing the evidence of their fulfillment of these goals, as well as providing examples or reflections of competence inhibitors.
If the application is compelling enough, the candidate will be invited for a 30 minute interview conducted by two independent assessors who are experienced in AI ethics assessments.
If the interviewers are convinced of your qualifications and the claims made in the "Case for AI Ethics competence" then you will receive the credentials of a IEEE CertifAIed assessor.
Your name will be entered into the list of certifAIEd assessors online if the candidate wishes to do so.
Becoming and assessor is completely optional and the candidate can choose to just pass the certification exam, which in itself is quite a worthwhile exercise.
I encourage you to read the IEEE Ontological specifications for transparency, accountability, privacy and algorithmic bias, irrespective of whether you choose to do the training and certification.
I'm happy to report that I have completed the program and am available to help enterprises with assessments of their AI use cases. Contact me via this contact form.
