AI Concerns

Tue, 06 Jan 2026 22:54:47 -0500

Its bad in new and exciting ways I never expected

This is what I kept saying at dinner tonight with my friend Andrew. We were out to dinner in between rounds of Undaunted Stalingrad and got on the topic of AI. It came up because our hometown favorite grocery store Wegmans stated collection biometric data in some of its stores which drew huge deserved backlash. I wanted to outline some of the concerns we discussed at that dinner.

AI Facial Recognition is Different

We are all used to cameras in stores and public areas in cities, and I am not entirely against them fully, although I do have reservations. Stores have a right to protect their goods and employees, and cameras help them identify criminals and security risks. If an incident occurs, they can go review the footage and hand it off the authorities.

When an incident happens.

Thats one of the big differences with AI facial recognition. It is constantly watching you. Before, there wasn’t someone in the back room checking how long you stared at that box of Fruit Loops before setting it back on the shelf. There wasn’t someone listening to who you were on the phone with. Or looking at who you went shopping with and linking you as friends in your file. But AI can do that, and that is different than just being recorded…

Now if I want to know everytime a certain person came to our store, I can list them all out. I could tell you how long they stay on average. Who they come in with. I can predict when they will be here again. These are all extremely problematic ideas.

First off - why do we give companies and cities the right to track us? This is not an inherent right that they have. Privacy needs to be legistlated yesterday. Walmart shouldn’t have the right to collect and sell information about me. That is my data, not Walmart’s, not Wegman’s, not Facebook’s.

Second - this is prone to abuse. Lets say someone working at a store with one of these systems is a stalker. All of a sudden, they have access to a potential victims routine and paterns. In this video researching Flock AI camera security vulnerabilities Ben does a great job going over the potential issues of this technology in the wrong hands - and to be clear it will fall into the wrong hands. If we allow this to become widspread, bad actors will use and abuse these systems.

What is the plus side?

Lets be honest, at best companies get slightly better lose prevention for this. If your local city or PD impliments it, maybe they are able to catch a few more petty crimials. But at what cost? Liberty lost is rarely taken back. Once we allow every private company and the government to track you with AI cameras, there is no taking it back.

Energy Implications

One part of the AI issue I rarely see discussed is the energy, land, and resource usage. Normally enviornmental impact is something towards the front of our minds with new technology, but for some reason the general public seems pretty quiet on the large amounts of energy, land, water, and resources used in the construction, build out, and operation of these data centers. Even today, while we are still fairly early in the AI boom, large tech companies are asking to revive long closed power plants to provide energy to their new AI focused datacenters. Not to mention the large amounts of land these massive data centers consume. The concrete to build them and the cooling systems used also consume massive amounts of water.

These datacenters are going to have massive enviornmental impacts. AI is by FAR the least efficent computing in the mainstream. By miles. Its not even close. Arguements between C and python are hilarious now that we are in the world of AI computing where we run a 100GB text model to tell you what to make for dinner tonight.

The Loss of Knowledge and Expertise

Forum usage has dropped to an all time low. People aren’t asking questions and having troubleshooting discussions where they can be indexed anymore.
This graph shows stack overflow submissions overtime. I did not validate the content of this graph, so please take it with a grain of salt. But I think we can all agree the internet has felt a lot more dead since AI started taking over.

On top of this, the jobs AI is automating today tend to be entry level jobs. It has, so far, mostly failed to replace senior technical resources (at least in IT|Cyber Securty).

This leads to a couple of problems:

Inablity to find information via web search

Have an issue that you are pretty sure is very common? Well too bad, the answer isn’t going to be on Google in a few years. If people are moving their troubleshooting conversations for technical issues to private, non-indexed AI tools search tools are going to continue to get worse (and lets be honest, they are pretty bad today). This is also an issue with Discord and Facebook groups over forums, but that is a rant I will save for another day.

A lack of juniors today is a lack of seniors tomorrow

It is looking like the tech industry is going to have a very similar issue to the issue the trades have had recently. Very few new workers are joining the workforce, and the current workers are aging out and retiring. This has lead to a shortage of workers in these industries causing increased prices and delayed timelines.

If we do not find a place for entry level workers in tech to sink in their teeth and learn then the gap between skilled seniors and the incoming juniors will continue to grow. This will lead to security issues as inexperienced workers are forced into possitions they are unqualified for with little oversight as the seniors get streched thin

A lack of ability to troubleshoot

One of the most important skills someone early in their tech career can learn is how to properly and quickly troubleshoot issues. Part of that workflow involves searching for similar issues so you do not have to reinvent the wheel. If the only spot you can find troubleshooting information is via an AI, then people will begin to fully rely on these tools and will lose many of the troubleshooting skills that are key to progressing their careers. I have already seen this trend in those entering the workforce right out of college - they more often than others will accept the output of an AI as the way things are “supposed” to work, instead of understanding that like any search tool, AI often supplies an incorrect answer or gives you a spotty explaination. You as the end user need to be able to digest that information and decide if that solution seems valid before taking it as gospel. I am going off on a tangent here, but hopefuly I made my point.

AI is coming. There is no stopping it. I wish there was.

Ai on CatalystND