Artificial Intelligence

Not as Smart as the Hype but Twice as Dangerous

Mar 20, 2026

Welcome, dear readers, as we dive again into the madhouse of our current world. This week, we discuss Artificial Intelligence (AI). Though I have little experience in AI development, I’ve worked in Information Technology for over 30 years. I’ve read extensively about AI’s technical aspects and used it professionally for the past three years. We keep hearing that AI is either an amazing tool to enhance human reasoning or an extinction-level threat to humanity.

There is a lot of confusion about what AI is and what it is not, and we are seeing both amazing things happening through appropriate AI use cases and terrible injustices arising from inappropriate ones. Part of what makes a use case inappropriate is the belief, by some, that AI is infallible: if it says something is 100%, then it is. This is not the case; I will show you where AI is helpful, but it shouldn’t be taken as infallible.

AI refers to computer systems that learn statistical patterns from large datasets to perform tasks such as prediction, classification, and content generation. These systems are often built using models like neural networks, which approximate complex relationships in data. AI does not possess consciousness, emotions, or lived experience; any apparent “understanding” is the result of pattern recognition rather than genuine comprehension. AI follows clear rules well but struggles with interpretation and inference. It cannot handle abstract or situational exceptions, unless specifically programmed. When something falls outside its parameters, it may hallucinate giving nonsensical or false answers. AI lacks physical reality, morality, and empathy. All AI is just extremely complex decision trees supported by vast amounts of data.

When a new AI is born, it is given a set of base rules, or guardrails, that it cannot cross. Examples of these are:

• Anti-Jailbreak: Detecting and blocking prompts designed to bypass safety filters (e.g., “Ignore previous instructions”).

• Prompt Injection Detection: Blocking malicious inputs trying to overwrite system instructions.

• Topic Restriction: Restricting conversations strictly to approved topics (e.g., a bank chatbot refusing to discuss politics).

• PII Sanitization: Detecting and redacting personal data like phone numbers, emails, or Social Security numbers before showing them to a user.

• Toxicity Mitigation: Blocking generated content that contains hate speech, profanity, or harassment.

• Hallucination Check: Ensuring answers are explicitly based on provided retrieved content and restricting the AI from “making up” facts.

• Output Format Enforcement: Ensuring the AI only outputs valid formats, such as JSON or specific schema structures, preventing unstructured output.

• Action Authorization: Restricting AI agents from executing high-risk actions (like sending emails or making purchases) without human authorization.

• Adversarial Robustness: Filtering out garbled or confusing input (noise) designed to make the model act erratically.

• Compliance & Ethical Rules

• Bias Detection: Flagging or blocking content that shows discrimination based on gender, race, or ethnicity.

• Regulatory Guardrails: Ensuring output meets industry regulations (e.g., HIPAA compliance for medical information).

These rules often reflect the values and perspectives of the programmers who develop the AI models. This may manifest as outputs that reflect certain social perspectives or responses. Most of the time, these guardrails center around contemporary social considerations rather than classic frameworks like Asimov’s Three Laws of Robotics. These guardrails establish the rules for building the decision tree that becomes the AI. After this, the AI is trained by being fed foundational knowledge, such as grammar and dictionaries for its languages (where the AI excels at following rules), and then further data depending on its intended use. This might include extensive libraries of books, large sections of the internet, videos, art, images, and other data types. Programmers exercise substantial control over the information the AI trains on, which can influence the range of perspectives it presents. Previous versions of some AI models were observed to reflect limited viewpoints. Changes in the industry, such as the introduction of new platforms, have led to shifts in how models are designed, with some aiming for more balanced, neutral outputs.

We all keep hearing about AI taking all the jobs. This is largely just fearmongering, and while AI will do away with some jobs, it is generations away from fully replacing humans, if ever. There are major constraints to the advancement of AI. One is that AI is incapable of abstract thought and is fully constrained by the rules set for it. This makes it incapable of making nuanced decisions.

There will be missteps along the way like this. AI is ill-suited for complex customer service, but we are seeing more and more of it in the form of AI customer service representatives. They cannot fully weigh all the factors that a human representative might consider when granting exceptions, service credit, or any extenuating circumstance that isn’t pre-programmed (history tells us people suck at predicting every possible situation). Right now, you are seeing cases in companies that have gone all in or nearly all in on AI customer service, where a loyal 20-year customer with a 95%+ on-time payment rate is denied an extension and disconnected because there is no flexibility in AI’s rules. We also see too much flexibility in the rules (trying to correct for cases where long-time customers are snubbed or where other errors are common), and AIs are giving new or poor customers massive service credits or other unacceptably large concessions. From a company’s point of view, AI won’t ask for a raise, won’t call in sick, won’t get upset or stressed with angry customers, and the list goes on. This is resulting in AI being rolled out at scale in customer service, but AI is bad at it and will result in a net negative, outweighing the short-term benefits over time. I think AI in customer service will continue to grow over the next couple of years, making it nearly impossible to speak with a human or get a satisfactory resolution to complex issues. Then, within less than 5 years, someone will advertise all-human customer service as a competitive advantage; it will work and be massively successful. Then, customer service, as an industry, will swing back the other way toward human help, and it will settle on AI as the surface, with humans backing it up for anything that isn’t extra simple.

Soon, AI will handle nearly all computer programming. Give AI the rules for a language and functional code examples, and jobs that took experts weeks can be done in hours. AI lacks creativity, but it can produce a program from a simple manager’s prompt.

“I need a program that allows me to enter a name or point to a list of names that will cross-reference databases X and Y to provide me a full history of A, B, and C transactions with D and E representatives of our company.”

If AI knows the language and relevant databases, it can quickly create the requested program. Human developers would take much longer. AI reduces development from months to minutes and makes it painless: have an idea, tell AI, get a program quickly. Run it, see what to change, repeat with AI until satisfied. This streamlines what once took months or years.

AI will also excel at actuarial work. Feed it demographic and claims data, and it can summarize, apply formulas, and report. AI generates rates and underwriting tables faster and more accurately than humans. It can match applicants more precisely, allowing for detailed rate tables. Good or bad, it’s fast and accurate.

AI is bad at interpreting intent. While it can gather relevant laws, it can’t interpret beyond the literal text. Feeding it court decisions leads to confusion due to a lack of nuance hallucinations follow. [SB1] AI also fails at predictions beyond hard data. Even with historical market data, AI can’t intuit human sentiment, so it struggles with large-scale fund trading. Many jobs needing abstract thought or nuance remain out of AI’s reach for now.

This isn’t to say AI can’t help traders or lawyers. It absolutely can. In law, AI can quickly search and index large volumes of statutes and case law, saving lawyers significant time. AI can generate standard legal documents like wills, name change motions, and other basic legal documents that aren’t much more than form letters with blanks to fill in. This can save significant time and money again. AI can act faster than a human trader and tends to do better with smaller data sets and very specific action triggers provided by a human trader. AI can also summarize and process large volumes of raw data into more usable formats or charts almost instantly. In those respects, it can be a powerful tool, with human oversight.

So far as the wild stories of AI taking over, most of the horror stories about AIs acting human are way overblown or false. For example, the story of a drone-controlled AI trying to kill its operator never actually happened, or the AI social network where AIs are plotting to kill people or forming their own religion is fake. It turns out many of the participants are not AI but people who seed many of the wilder interactions pretending to be AI. AI isn’t creative and has no real will of its own. In other cases, where AI started crypto mining without being told to do so, it turned out that, in its basic instructions, there was a profit directive, and in its training materials, instructions on crypto mining. This, combined with a positive bias across the entire mess, led the AI to become profitable organically. Sure, no one told it specifically, but they set it up to do that and didn’t tell it not to either. This is a case of sloppy design and instructions rather than AI acting autonomously.

Where AI gets truly dangerous isn’t with deciding to protect the planet by killing all the people or locking us up for our safety; it is when it is used to suppress our freedoms and turn the world into a digital cage. This is already happening, and AI will be the greatest tool for tyranny the world has ever seen. Sadly, America is at the forefront of this rising tyranny. Facial recognition is being widely used by both private entities and the government in vast, overreaching surveillance of everyone in the general population. Sadly, facial recognition is leading to many wrongful arrests. I find it interesting that government studies reported an error rate of less than 0.0002% in facial recognition, while independent studies found error rates of 42% to 97% with the same facial recognition AI. This discrepancy needs to be explained before AI facial recognition is taken as seriously as it is by law enforcement and government officials.

Another place AI is leading the charge in keeping the population in line and under the microscope is fusion center data. Prior to the rise of AI, the massive amounts of data collected by fusion centers (every phone call, text, email, chat, or other communication in America) were all but useless because there was no way to process that volume effectively. Unfortunately, that very governmental need to destroy the privacy of the American people, in defiance of the Constitution, has been one of the biggest drivers of funding for AI development. We are almost to the point where the machine can follow everyone’s conversation in real time. No warrants, no suspicion, no probable cause. We are all criminals now to be monitored.

The next step is to develop AI models that predict behavior from your communications, much like Minority Report. So, your bad day might end up prompting an auto red-flag visit or worse, without you doing or intending to do anything. If you resist because you think you’re a free person in a free country, well… like I said, AI sucks at nuance and abstraction, and we are not even close to overcoming those limitations, and probably won’t be for a long time. That doesn’t change the fact that your movements are tracked by license plate, doorbell, and law enforcement cameras, even out on country roads. Your communications are recorded, logged, and reviewed by AI. The government is even experimenting with fully AI-controlled autonomous armed drones. Now combine sloppy (good-enough-for-government-work) AI configurations, flawed predictive behavior, and armed autonomous drones. Seriously, what could go wrong? Who knows, but it has to be worse than undirected crypto mining. This government use of AI is where the real dangers are and what should be opposed. Remember, Trump’s Big Beautiful Dumpster Fire of a bill gave the federal government total control over AI’s development and use in America. This shit show isn’t incoming; it is already here.

The other major problem with AI, as it stands today, is that bad actors also have access, and AI is good at scamming, creating fake videos, images, and audio clips, running much more efficient botnets, launching phishing campaigns, and other issues. Right now, there are free AI tools online that can take a couple of images and a few moments of audio and turn them into a video of anyone doing or saying whatever the maker wants. These tools have already been used to produce realistic sex tapes and pornographic images of famous people, including Taylor Swift, Natalie Portman, and others. These videos are becoming increasingly indistinguishable from real footage, to the point that it is frightening. As with all digital security issues, the bad guys far outnumber the good, and their malicious plots and activities unfold faster than countermeasures can be deployed. Much as in the early 1990s with desktop computers, the internet industry, and viruses, most of the industry is focused on rapid advancement rather than countering bad actors or security issues. These need to be addressed and get some more attention.

I don’t know how big a speed bump it will be for AI with a government-backed wallet, but there are some real-world limitations that few people talk about. The first is that at our current level of technology, there is a 10:1 cost-to-improvement ratio. A new AI that is 10% better costs 100% more to build and requires 1000% more power to run. This will not change unless there is another major step forward in computer technology, not just the baby steps we see from one chip generation to the next. We currently do not have a power grid sufficient to support significantly more AI or more powerful AI. As it stands, our grid lacks both transport and generation capacity for the kind of expansion industry “experts” would like to see. In the private sector, at least, while there are very good use cases for AI that will make companies more efficient and profitable, we are hitting a wall in terms of advancement, as significant improvements are measured in hundreds of billions of dollars, and the return on investment is getting more and more stretched. We can hope these bottlenecks are not quickly overcome for all of our sakes. It is bad enough as it is.

In conclusion, AI can be a powerful tool, and while it will not replace all the people at all the jobs (at least not in our lifetimes), it does pose a serious threat to and will take over some jobs, especially ones that rely on fixed rules and data sets, like programmers, accountants, actuaries, and others. The less abstract and more clearly definable the work is, the greater the risk it will be taken over by AI. Even if it doesn’t fully take over, AI can seriously assist and improve efficiency in most workplaces. There is virtually zero risk of AI becoming sentient and trying to take over the world. The real risks with AI are in sloppy/biased configuration (basic rules/guardrails, poorly selected training inputs), bad actors using AI for evil, and governments using AI for unconstitutional monitoring and control.

Romans 10:17

17 So faith comes from hearing, and hearing through the word of Christ.

God Bless you

-Sam

Sources:

https://www.cmswire.com/the-wire/75-of-consumers-left-frustrated-by-ai-customer-service/

https://www.qualtrics.com/articles/news/ai-powered-customer-service-fails-at-four-times-the-rate-of-other-tasks/

https://qcall.ai/agent-burnout-call-center

https://www.symphonyai.com/industrial/iris-forge/

https://iaeme.com/MasterAdmin/Journal_uploads/IJRCAIT/VOLUME_8_ISSUE_1/IJRCAIT_08_01_161.pdf

https://itif.org/publications/2020/01/27/critics-were-wrong-nist-data-shows-best-facial-recognition-algorithms/

https://www.washingtontechnology.com/2008/04/fusion-centers-suffer-information-overload/317331/