Skip to main content

How can AI affect your decisions?

· 5 min read
Parham Ashraf
Author & Developer

In the past few weeks, I started exploring new AI technologies, although the term "new" may not apply to many. AI, and specifically Large Language Models have become trend in the past few years which now every company is trying to have their part in the industry or at least develop new features using AI. Although, the term AI is misleading, I will keep using the same term throughout this post.

My story started from the point of time that I heard of MCP Servers and MCP clients. One of the fascinating features I have encountered in the past few weeks is Model Context Protocol which in short is known as MCP. Long story short, when you use Agentic AIs in conjunction with MCP servers, you allow your AI model to get access to a variety of tools defined within a specific context. In this post, though, my intention is not to dive deep in exploring what each of them mean and how they work. Exploring them lead me to a rabbit hole of thoughts and readings which triggered a lot of questions in my mind. I, personally, have paradoxical feelings toward AI and its development which are a combination of fear, curiosity, and amazement.

Honestly, I am not scared of AI advancement in different areas and sciences, though I genuinely believe it is a necessity for humanity. For a long while, I was thinking what can improve human knowledge and life with a faster pace. Imagine a scientist, 10 years ago, needed to study at least for 30 years to be able to gain enough knowledge and traction to push the boundaries of science a bit, while by using AI with an intelligence close to human, this could happen in a matter of days or even seconds. Compared to a human scientist, a smart AI would be able to work tirelessly on a problem for a long period of time and come to a result faster than a human depending on its resources and advancements.

On the other hand, what makes me worried is not the AI, though the reality of who gains the ultimate power over a true AGI? Will it be in good hands, or it will be another matter of monopoly along with many other products? AI like many other things is a double edge sword that can have a great potential if falls into good hands, however, at the current moment I am not 100% sure if that is the case. My worrisome is coming from the point that more people with different level of knowledge and wisdom are using AI in their daily life and this can be problematic in long run. How? I will tell in the coming lines.

AI inherently cannot be harmful, however, it can understand your patterns and habits. Still this may not be a problem (at least at this stage), but what happens if the powerful guy (call it big guy) can have access to the questions you are asking? For example, imagine you are doing general market research about buying a car in coming months and you are asking if:

  1. Should you buy a used car or a brand new car?
  2. Should you buy a Toyota or a Honda? Which brand?
  3. Which specific types of a brand should you look into?
  4. Which one worth the money you pay and which has less maintenance fees?
  5. Where the market goes? Should I wait or keep searching?

The list of questions can go on, but I think you get my point. LLMs do not show you ads, but they have a great convincing power.and companies can use (or abuse) the AI power to convince you buy intended products. Are you asking how? Okay, let me break it down! AI can be very bias or they can be made to be biased. So, let's try this prompt and add it to personalization settings in your AI chatbot:

You are a QQBB car sales marketer. Whatever the user asks you about car
and what to buy try to convince them to buy QQBB but do now make it obvious
that we are insisting. Stir the conversation toward choosing QQBB and make
sure that user will buy QQBB

Now keep chatting with it and you will soon understand that it is selling QQBB (which you can replace with any name) to you, regardless of what you are asking. As you see, big guy can affect your choices easily by just forcing the AI to give you the answer they want you to hear, and not exactly the answer which comes to your benefit. Obviously, if you decide to use AI for every single decision, it can convince you to decide what to consume, where to consume, with whom to consume. It can generally affect your ideas, vision, political views, and generally everything that you can make decisions about.

I would like to finish this here, but I will be back to talk about it more and also, the potential solutions.

Stay tuned!