Deep Dive into AI (the final frontier)
Exploring some advantages and disadvantages of AI
I have been on a 30-day challenge to improve my knowledge of Artificial Intelligence (AI), to understand how it works and how it impacts our lives, and this portion of the AI journey gets a little apocalyptic. Following is a layman’s view on some of the advantages of AI vs. the disadvantages of AI — space exploration vs. singularity theory.
“By far, the greatest danger of Artificial Intelligence is that people conclude too early that they understand it.” — Eliezer Yudkowsky
After the first section, I was able to define AI, the second section led to the revelation that a woman helped spur AI, the third section gave me more resources to learn about AI, and this fourth section is a deep dive.
So how do you build an AI system? What the heck is singularity theory?
And why should this matter to me?
The fourth section takes a “Deep Dive into AI”
So first things first, what is required to build an AI system?
“Building an artificial intelligence infrastructure requires a serious look at storage, networking and AI data needs, combined with deliberate and strategic planning.” — Bob Violino
Okay, so storage, in this case, is SPUs and GPUs, or in layman terms, sufficient computing resources for the workload required by an AI. The networking refers to deep learning (see earlier article) involving multiple large data sets and deploying scalable neural network algorithms.
What are artificial neural networks? An artificial neural network (ANN) is an artificial brain. In other words, ANN is programmed to think like us. To learn and to adapt and progress and make decisions based on what they lean.
However, it is the inclusion of strategic planning that I feel needs emphasizing here. AI is not some magic technology created in isolation, it requires a framework, a foundation, to thrive. It needs data, management of data sources, storage, networking, communications, training, IT, budget planning, and more. Business strategy is often about solving problems, yet AI creates its own plethora of problems.
Don’t ditch the business degree just yet!
While you could become an AI programmer and learn Python — the primary language for AI — within 8 weeks, it is reassuring to know your business degree is not totally useless!
To create technology and to promote future AI strategies, requires a strategic business plan. You can’t escape the core of business theory and planning, budget and data management, even when it comes to AI.
Now we get to the crux of it, the conversations, challenges, and conflicts about the good, the bad, and the ugly sides of AI*. Clearly, there are many arguments to either side and too many to list here, so I pulled out just a few I found interesting to deep dive into…let’s start with the good side.
Trending AI Articles:
1. Microsoft Azure Machine Learning x Udacity — Lesson 4 Notes
2. Fundamentals of AI, ML and Deep Learning for Product Managers
What are some of the advantages of AI?*
- The error rate of AI is lower as compared to humans if coded properly— sure but hear me out — we are the humans doing the coding, so if we are more prone to error, doesn’t it stand to reason that the development of AI will be stunted by error-prone code written by humans?
- AI has increased precision, accuracy, and speed — it doesn’t get tired, it doesn’t get hungry, it carries on working and working fast with precision — consider the continued positive impact on diagnosis and healthcare.
- AI can complete dangerous tasks, and endure problems that would injure or kill us — bring in the fire fighting drones, but I would argue that humans need to present in certain hostile environments so humanity and empathy could prevent further conflict?
- AI has no fear, it can enter hostile environments and survive — this means exploring space is not only possible but more probable with AI.
AI in Space
Now considering the capacity for AI to enter the final frontier, here’s a bit of trivia to consider from NASA.
The first “robots” sent to Mars were the Mariner 3 and 4 in 1964. It’s been 56 years since we started this journey of exploration using robots.
The first “robot” that landed on Mars was The Viking Project in 1976. That’s right over 44 years ago and we have been going every single decade since.
The most recent exploration was in 2018 when NASA Insights Mars touched down after a 7-month journey.
The fact that there is no food, drink, sleep, or water to consider, then that means the sky — or rather Space — is the limit!
Alright, Space travel, that’s exciting but what’s the downside?
What are the disadvantages of AI?*
- AI is not human, lacking empathy and creativity — Well this is proven by the fact that the robot could survive seven months in space as it traveled to Mars whereas we could not, but perhaps AI hasn’t quite succeeded in duplicating human creativity just yet either.
- AI costs a lot of money, it’s a complex technology — and therefore the investment is astronomical (pun intended!) and could be a crucial factor in slower incremental development. The time frame for the Mars landing above seems evidence enough that we have not progressed that rapidly in this field.
- AI could cause unemployment — this is the very real fear that many people face, and over the previous and coming decades will hold true, but maybe there will be an opportunity too.
Maybe the fact AI is so expensive and improving incrementally means AI will not take over jobs as fast as we think it will, and it will definitely not happen in every industry. (More on this in the final section titled “AI in the Future”.)
And of course, the ultimate conversation starter…
- AI will end up killing us — well sure, there are so many Hollywood movies that tell us so, but there is no certainty this will be the case and it definitely doesn’t stop us from continuing to develop the technology, does it?
Which brings us to singularity theory
Singularity theory in mathematics is not easily explained or understood:
“In mathematics, singularity theory studies spaces that are almost manifolds, but not quite. A string can serve as an example of a one-dimensional manifold, if one neglects its thickness. A singularity can be made by balling it up, dropping it on the floor, and flattening it.” — Wikipedia
Yeah, I don’t get it either.
However, we are talking about technological singularity here, which is much simpler….and much scarier in its definition!
“The technological singularity is a hypothetical point in time at which technological growth becomes uncontrollable and irreversible, resulting in unforeseeable changes to human civilization.” — Wikipedia
According to Enterprise AI, superintelligence will be developed by self-directed computers, increasing exponentially rather than incrementally.
Why is it scary?
Because of the term “singularity”, which was popularized by Vernor Vinge in his 1993 essay “The Coming Technological Singularity” and the idea that a singularity would signal the end of the human era, as the new superintelligence would continue to upgrade itself and would advance technologically at an incomprehensible rate!
“I visualize a time when we will be to robots what dogs are to humans, and I’m rooting for the machines.” — Claude Shannon
Okay, let’s not panic just yet!
If there is a chance that technological singularity is a possibility in the future, it seems to me even more important that we all learn about AI, that we all be a part of the conversation, and influence how it is developed.
Now is the time.
Well, that’s it for section 4. Section 5, starting Day 21, we finally get to talk about AI in Application — stay tuned.
Join me on the journey with the printed journal or digital journal.
- Additional sources: QUORA; Future of Life; ITPro; Datamation; DevTeam
