Thoughts on AI Index Report 2022 (Stanford)

This is a continuation post (Part 1 here), on my thoughts on the AI Index Report found here.

The rise of AI ethics everywhere

I always believe to do good research in AI Ethics, having many use cases is vital, with each use case comes more lessons in implementing AI ethically. The highlighted fact that researchers with industry affiliations contributed 71% more publications year over year, based on the report, substantiate my viewpoint. In order to build up that knowledge base of use cases, we need more talents to do with the implementation and again you can see the bottleneck is talents.

I do not believe in the alarmist brand of rogue AI, however, I feel we should start looking into it right now since we are at the building and research stage. It is very important for all of us to pay attention to building AI that provides positive impact.

Notice I do not say AI Ethics is about making AI ethical, rather it is about making AI designers and professionals more ethical rather, see my post here. https://lnkd.in/gANVha4Z

AI becomes more affordable and higher performing

Like any technology, it will just become smaller, more compact, and more affordable as time goes by. Just look at our disk storage and mobile phones (now smartphones). The two questions right now is:

a) As a company, are you going to jump in to build your capabilities, or are you waiting for things to get easier and more affordable? Remember, you are competing in the industry, you will lose the lead if you wait, HOPING your competitor does like-wise.

b) How can you build functional, scalability, and flexible infrastructure to take advantage of this trend then?

Data, Data, Data

Data growth is given, especially with technology being adopted in more industries. The key question is how to take advantage of it. The very first step will be on how to manage the rapidly growing data assets, and this is where data governance, and management come into place. If you have been following my posts, you will have known that I strongly advocate good quality and well-managed data first before all these mambo-jumbo Machine Learning sold by bootcamps and study programs. After that, before going into the Machine Learning model training, will be the managing the labels, something that Andrew Ng has been advocating for more than a year, after the previous research era was more focused on building better algorithms.

Data collection needs to be planned in order go get good quality data and data need to be well-managed to maintain the quality before building your Data Science or Artificial Intelligence capabilities. To do that, your organization will need someone experienced rather as data collection, governance and management is not taught in a lot of bootcamps or study programs based on observations.

What are your thoughts? Will love to hear from you! You can share them with me on my LinkedIn. Please feel free to link up on LinkedIn or Twitter (@PSkoo). Do consider signing up for my newsletter too. Will continue my thoughts on the other points made in a later post. :)