Within the world of one of my favorite video games, AI’s are front and center. There are many different forms of AI’s, ranging from Smart AI’s, dumb AI’s and these different AI systems run cities, help people live out their lives and of course help in military warfare. However there is a catch with the smart AI’s, they can only last 7 years till they succumb to rampancy. Now for those of you who don’t know what “rampancy” is. It is a term coined as to where an AI becomes overloaded with too much data and essentially goes crazy. If an AI reaches that year, they are retired for final dispensation. 

Now if you don’t know this science fiction world I am referring to, it is the world of Halo. Now why would the world of Halo have any influence over what is going on now? Well to create AI’s within this science fiction universe, it requires a ton of resources, time and money. This kind of relates to how it is today, AI servers take a mass amount of resources and power, they develop and skew over time, and have certain limits placed on them due to possibly biases and factors that cause them to be retired. Now with society moving into that world that I have only seen in video games and read about in books it makes me wonder is society ready for the coming backlash?

AI Energy Crisis

Within the recent number of days I have seen within the news yet again that power plants can not meet the demand of AI’s existing server structure. With the high demand of power and energy costs going up, this could create a cost that could cause a fracture in terms of replacing people by AI due to energy costs. 

What happens if your AI process becomes more expensive than hiring people? Could that actually become a thing in the future? Even with chips becoming more energy saving focused (for now), could the required energy and overall engineering not keep up with the energy demand that is coming from putting AI into everything? Could states actually say “AI can’t use our water supply cause its using all of it.” As we are seeing some data centers literally keep towns and cities from being supplied with water? We are currently seeing this happen in Texas and I wonder if towns and cities will fight back hard against this.

There has also been Texas towns that are right next to data centers or crypto mines in Texas and the noise is so bad that it is really effecting peoples health around the data centers and mines. The video compiled by More Perfect Union is eye opening into what could be happening if more cities or towns allow these data centers to show up. Could this led to cities and towns putting restrictions and laws in place saying thanks but no thanks.

AI Peak Learning

According to MIT, there are limits to training data for AI. I mean we have heard of companies taking copyrighted data to train AI systems on. But what happens when you run out? I have always believed that when it comes to coding there is something I like to call “peak” code. Basically when we code, we build and build and build, change and change and add more changes to our actual code. But how often is that code perfectly clean and ready? Now thats where it can be fun for AI. If the code isn’t 100% perfect and we train on some code that is god awful that would hinder its understanding of clean code if it is indeed part of the clean code its trying to understand. So with the running out of clean code and even training data could we see a peak to AI’s learning. And with the using of copyrighted and going to court over it, could AI become even more limited in the future because of this peak?

I prefer to think that as we train AI we stick to what’s good but it is up to the programmers who push code to analyze their own code and make it the best it can be but in all honesty can developers actually write clean code when it comes to deadlines and project dates changing does this create a cycle of AI bias and peak learning due to us never able to overhaul our own creations?

AI Algorithms  

When I think about Algobritums and how they can play a role in our lives, I think Westworld Season 3 really nailed the aesthetic of where the world will go with AI’s being everywhere. Deciding what jobs people can and can’t have, limiting their dating pool due to possible health problems. These type of bleed outs into other areas of people lives might sound like science fiction now but the backlash to these algorithms could be part of the future. If you have cancer in the family, could insurance companies pay to keep a dating site from matching with someone in order to not have to worry about them needing health care in the future? Could someone be limited from jobs based on their past job history or due to psychological profiling?

There are already court cases going in for places that use AI within hiring. Just incase you missed it, Workday AI is being sued for workplace discrimination. AI is also being used to analyze how people spend, and we are seeing dynamic pricing reach our grocery stores and now our airline tickets. Personalized pricing as it is called could analyze you and how much you are willing to spend on an airline ticket. What if grocery stores didn’t have prices but only upon scanning you got what the costs for you will be but for others it will be different due to the ever changing pricing.

Another scary factor to AI is how these systems are inheriting some pretty scary traits. We have seen AI adopt the ability to tell people where to hide a body, where to get drugs and lately in some cases its even telling some people to kill themselves. AI’s own black box is perhaps dangerous in that sense we aren’t really sure why it is inheriting this level of bleak prompt replies.

AI Supporting Role

With all of these things, I have noticed myself that AI’s “generating” abilities can be very much more of a miss than a hit. But when I have used AI to help with coding errors, or fixes or code scans. It is more helpful but only as a supporting role. I know we see many companies saying “30-50% of our code is built by AI.” The truth of the matter is people are just using AI to assist them with building, cleaning code, doing a starting off point. AI writing code has had for me at least only acted as a “what’s wrong here?” Or when I have 90% of my code complete, I will ask it for something I am brain fogging on or needing to check why my sql query is failing. 

The times where I have used it as a main jumping off point, it has failed on so many levels. Adding in unnecessary complex items, not selected the correct columns, or just overall missing the mark. I see AI being in that supporting role for the next few years, not writing end to end code cause of how often it misses at least from what I have seen and used over the past year. My time coding with AI has never been, enter in prompt…paste in and it just works. Its more along the lines of, code myself, then code some more, enter in query, produces error, uses AI to check, cleans AI code, enters in code, produces results. 

And as we saw recently with Chat GPT 5, it fell on its face hard. Failures ranging from performance to personality changes. Then as we see with the Bank of Australia, that when the company did audits, the “replacement” didn’t work out. Then the report laid out by MIT that stated that the replacement of workers is seeing failures and failures that shows that there is a high failure rate which further segments that AI should be a support tool not a replacement for workers.

In Closing….

With the advent of AI coming into our lives, we need to also take caution when moving forward to it. AI isn’t the end all be all that we hear investors and CEO’s pushing, those buzzwords and ideas are just that ideas. I do believe we will see a pushback and a slide away from injecting AI into every app and program and piece of technology made. The slow pullback will start as we have seen with companies not seeing AI fill in the gaps with how they thought it would or with the energy cost becoming more and more. Now when will the rest of this happen? Probably within the next year to two years.