Artificial Intelligence (AI) and Machine Learning (ML) have been around for several decades, initially accessible mainly to researchers and large corporations. However, advancements in technology have made AI and ML tools available to everyone, from small businesses to independent developers. There’s no question that they’re here to stay. As a developer, I generally embrace new technology and get excited about the benefits both personally and professionally. The advances in AI and the speed at which they are progressing have made me extremely excited, but admittedly, they have made me very nervous as well.

But why?

The obvious reasons include job loss, or the fear that my step-children, nieces, and nephews won't find work because machines do it faster and cheaper. Or a colleague being let go because a large language model (LLM) is actually quite good at generating copy or crunching numbers and running complex analyses. These scenarios play out in the back of my mind from time to time (I worry a lot), and while I worry about these things, I feel like this isn’t going to happen in the very near future… or will it?

Here at GO, we have embraced AI as a tool to improve our efficiencies and those of our clients as well. We currently use AI in a variety of ways in our daily work. For example, we use it to help with market research and competitive analysis, analytics, copywriting, retrieval augmented generation (RAG) systems, search engine optimization, code generation, storyboarding, and more. AI has personally saved me hundreds of hours already, performing monotonous tasks that I hate doing in record time. Programmatically, many complex tasks are now trivial.

For example, a client recently requested a seemingly simple feature in a secure portal we built where all parts of a street address need to be extracted from a single line of text (a string, in nerd speak). So, for example, consider this line of text:

2074 Park Street, Suite 307, Hartford CT, 06106

Automatically extracting the street, suite (if applicable), city, state, and zip code to be used as individual chunks of data might sound pretty basic but is actually very complicated. Remember, computers like consistency and patterns, but people can type a street address in many different ways. Prior to AI, you’d probably try to leverage the Google Places API to perform an operation like this. I’ll spare you the technical mumbo jumbo as to why this is difficult, but with AI, it no longer is. With a simple prompt and an API call to your favorite LLM, this is now easy. That’s a win in my book!

From an AI perspective, while helpful, this is pretty basic stuff. AI is getting smarter, though. There have been comparisons between the current state of AI and human developmental stages, describing AI as having the intelligence of a teenager or college student but predicting that it will soon reach the level of a Ph.D. scientist.

But none of this is why I worry.

I worry about Artificial General Intelligence (AGI). In case you haven’t heard the term, it is defined as:
“Artificial General Intelligence (AGI) refers to a type of artificial intelligence that possesses the ability to understand, learn, and apply knowledge across a wide range of tasks at a level comparable to human intelligence. Unlike narrow AI, which is designed for specific tasks, AGI can perform any intellectual task that a human can, demonstrating general cognitive capabilities.”

Think hard about the last part of that pull quote—“demonstrating general cognitive capabilities.” Don’t get me wrong, AGI is exciting, and I can see some amazing opportunities, but I imagine some frightening scenarios as well. For example, would you rather visit a doctor who has looked at 10,000 CAT scans throughout the course of their career, or use an AGI system that has seen 10,000,000? We’re already seeing AI being used to detect early-stage cancer that many doctors would have missed. This is amazing. But what about having a digital pet with AGI that lives forever, never ages, and is plugged in or put “to sleep” while you go on that fancy family vacation. Bizarre, right? If machines will have “cognitive capabilities,” then it’s not really a fallacy to imagine we might not ever have to go to a “real” doctor again, or a lawyer, a real estate agent, accountant, or even hire child care. Imagine the military implementations of AGI? Now that’s scary!

There is significant debate among experts as to when AGI will be widespread. Some experts speculate within five to ten years, while others say fifty or more. Either way, it’s coming.

The evolution of AI and ML has transformed from being niche technologies accessible only to researchers and large corporations to tools available to everyone, including small businesses and independent developers. This transformation brings with it both excitement and apprehension. While the practical benefits of AI, such as increased efficiency and the automation of mundane tasks, are undeniable, the potential implications of AGI are both thrilling and concerning. AGI's ability to perform any intellectual task a human can raises profound questions about the future of work, ethics, and society at large. As we stand on the brink of this new era, it is crucial to balance our enthusiasm for the potential of AI with a thoughtful consideration of its broader impacts, ensuring that we harness these technologies in ways that benefit humanity as a whole. I leave you with this question: How do you believe AI/AGI will shape our world in the next decade, and what responsibilities do we have in guiding its development?