We can tend to be late adopters. If you have every sold anything that is “bleeding edge” and been bitten by it still having bugs, or not yet truly available, you learn to give things a little time to mature before fully adopting them. But that doesn’t mean that we are not using AI for ourselves and our customers, just that we are being careful about it.
Take the above image, we have been using AI to generate fun images for our newsletter for about a year now. Actually, we have been using AI in machine vision for close to a decade now. That was before AI was such a buzzword. And you could argue that AI in machine vision is actually machine learning, not AI. But that is all part of what makes AI so confusing.
Another way we are using AI is in our own code writing efforts. We have three full-time developers in our group, and while that may seem like a lot for a business of our size, most of the time, it is not enough. So, using AI in coding to allow us to take on bigger projects, and complete them faster, for our customers is super appealing. Like AI in general, we do not plan to use it to replace people, but rather to make the people that we have better and more productive. Getting more work done for our customers for the same, or even less cost.
To us, the most incredible thing that we have been able to do with AI is customizations to our own ERP, Microsoft Dynamics BC. Doing work that we don’t have the expertise for in-house, but are able to accomplish without hiring someone with the needed skills because we can engage AI to do the work.
Being around as long as we have also means that we have some very old code base in our customers’ operations that will need to be updated to a newer version. We are now exploring using AI to accomplish that task with very little actual labor by our group.
Finishing with my favorite AI trick of all. What happens if you ask AI to comment about itself. Enjoy the below, which is surprisingly enlightening commentary.
AI Overview
Yes, the term “artificial intelligence” (AI) is widely considered to be misused, overhyped, and diluted in modern usage. While technically referring to machines that simulate human intelligence to learn and solve problems, it is frequently applied as a marketing buzzword for simple automation, basic algorithms, or traditional software.
Here is a breakdown of how and why the term is being misused:
1. Marketing “AI-Washing”
Companies often label products as “AI-powered” to appear cutting-edge, attract investors, or boost sales, even when the technology is minimal or non-existent.
- Examples: Smart home appliances, basic chatbots, or simple recommendation engines are often branded as AI, despite operating on predefined rules rather than true machine learning.
- Consequence: The term is becoming synonymous with “tech” rather than specifically “intelligent” software.
2. Conflation with Automation
There is a common, incorrect equating of automation with intelligence.
- Automation: Predefined rules and workflows (e.g., if-then rules).
- AI: Adaptive learning, reasoning, and decision-making based on data.
3. Confusion Between AI and Machine Learning (ML)
While machine learning is a subset of AI, the two are not always synonymous. Labeling every algorithm or data analysis technique as “AI” obscures the distinction between systems that truly learn and those that are just running advanced statistical calculations.
4. Overuse in Video Games and Software
Historically, AI was used to describe NPC (non-player character) behaviors in games. However, in the wake of generative AI (like ChatGPT), this traditional, “hand-crafted” logic is often conflated with “new” AI, causing confusion in both communities.
5. Anthropomorphism (Treating Tools as Human)
Using the term “AI” often leads to treating software as a sentient entity rather than a tool. Phrases like “an AI thinks” or “the AI believes” mislead the public into believing these models have consciousness, wills, or desires, when they are primarily pattern-matching probabilistic models.
6. Misuse of Terms Like “Hallucination”
What is popularly called an AI “hallucination” is actually an error in a probabilistic model. Using the term “hallucination” implies a mind capable of perceiving reality incorrectly, rather than a system making a mistake in prediction.
Why This Misuse Matters
- Erosion of Trust: When overhyped products underdeliver, it leads to disillusionment, skepticism, and frustration.
- Misleading Consumers: Users may believe a system is more capable, trustworthy, or “intelligent” than it actually is.
- Hiding Human Labor: It often masks the significant amount of human labor (data labeling, reviewing) that makes the “automated” systems functional.
Many experts suggest using more precise terms like “machine learning,” “automated decision system,” or “predictive algorithm” to describe these technologies accurately.
So, if you are looking for an automation partner who is cautiously embracing AI, www.emptechgroup.com is the best way to reach us.