Now I want to preface this post with; I am not an expert on this topic; I just find it fascinating. 🙂

I have recently finished reading “Human + Machine, Reimagining work in the age of AI”. It is a hot topic at present, given the advances that are being made in robotics and artificial intelligence.

This post is more of a collection of tidbits of interesting things which may prompt further research.

Universal Base Income (UBI)

Sinead Bovell is a futurist I have found on TikTok, and she speaks a lot about AI and its effect on the world.

One of the future states could be that technology surpasses human ability and starts to replace large portions of the workforce. In this event, there would need to be something that supports people’s lives, here enters Universal Base Income.

Essentially it is a globally set base income that people would get as a salary replacement. Now you could look at this from the dark side of things, but there is also a brighter side. You could now choose to pursue your path, learning and growth, music, art etc. and be supported in that endeavour.

How many times have you heard about musicians or artists that struggle through life just making ends meet because they want to do what they are passionate about, this would enable them to do that. UBI could help them realise this dream.

Empathetic Technology

Chris Milk is a filmmaker that did a TED talk on how virtual reality can be used to connect more people. His company work to create VR content that tries and convey the struggle of millions in the world, bringing that struggle to the doorstep of the United Nations; his goal? to bring understanding to those in ivory towers making decisions affecting millions of people.

Another rapid change in the world is using AI in human interactions. How many times have you been on a website in the last few months trying to find something and been directed to a chatbot. That chatbot, in this day and age, is more than likely to be an AI backend bot. Amazon’s Alexa and Apple’s Siri are two more examples of this technology.

What backs this technology is a Machine Learning technique called Natural Language Processing. As the name suggests, this technique takes natural language, processes it, and tries to understand the intent. “Do you have this shoe in another size?” or “What cable do I need to connect X and Y devices?”. The AI bot would then direct the person to the relevant answer.

What would happen if someone said to a chatbot, “Im not sure I can cope anymore; I have thought about taking my own life…”. Without empathy, how would a machine know it has a potentially fragile situation and person on the other side, and there is an opportunity for it to help. Here enters Koko, a company that brings Psychology, AI and Natural Language Processing together to help “train” machines to be more aware of the human side of its role.

Feature engineering

One of the most challenging parts of working with AI and ML is getting access to, gathering, curating/refining and prepping data ready to be used with this technology.

I embarked on 100 days of Machine Learning a couple of years ago. The aim was twofold, to get exposure and learnings in machine learning but, more importantly, to gain an appreciation for what it took to get something up and running.

There is a branch of engineering in data science called Feature Engineering. Simply put is the process of enriching a dataset with more data and then transforming that data into a better “shape” for machine learning to use it. A large part of this process was forming a “master data” set that contained all the data I needed to train and validate the machine learning model.

Binning is one method where you take a continuous series or numerical variable and create “bins” or ranges. This helps to simplify numerical data and looks to find correlations in groupings.

Another method is called Feature Encoding, and this is where you take a category column and transform it into numbers:

[male, female] → [0,1]

This is called label encoding.

or if there is a relationship between the information, given a list of priorities:

[low, medium, high] → [0,1,2]

ordinal encoding can be used, and the order can depict the “weighting” of the categories

A really interesting term that I read about in Humans + Machines was Data Exhaust:

Data exhaust refers to the data generated as trails or information byproducts resulting from all digital or online activities.

from here

This is where unstructured data lakes come into their own, being able to “dump” vast amounts of different data then enriching datasets. It is a powerful way to continually improve the accuracy and try to gain more insights using this “exhaust”, remembering ML is used to see patterns where we can’t.

Machine Relations (MR)

This was such an interesting topic and one that I will do more reach in. As AI and ML become more prevalent in society, there will need to be checks and balances put in place to ensure that the technology is doing its job.

Here enters the role of a machine relations managers, who will be similar to human resources managers except they will oversee AI systems instead of human workers.

They will be responsible for regularly conducting performance reviews of a company’s AI systems. They will promote those systems that perform well, replicating variants and deploying them to other parts of the organization. Those systems with poor performance will be demoted and possibly decommissioned.

A big part of this is to ensure the AL/ML is adding the value it needs to, is continually getting more accurate and above all else, is performing its job in the most ethical and unbiased way it can.

The future

AI, ML, VR and AR will have broad reaching impact on the world. Having a good high level understanding of these areas is virtual going into the future, even if you are not a purely technical person. The one thing (at present anyway) that can’t be replaced by machines is … the human part of the equation. This technology will augment everything we do and used well, will make the world a better, more interesting place.

Yours in learning.