Artificial intelligence (AI) captured everyone’s attention in the last 12 months. However, with that attention has come significant concerns about the spread of misinformation, the impact of AI tools on jobs, and even the existential threat to humanity.
We keep reading opinion pieces outlining doomsday scenarios about the potential for uncontrollable machine intelligence to wipe out humanity, but I believe the mystique around the development of artificial general intelligence (AGI) has been leveraged to get our attention.
Just to level set the conversation, AGI doesn’t actually exist today. When it does exist, it will be an autonomous system that surpasses human capabilities in most economically valuable tasks. It will be a form of AI that is just as cerebrally capable as human beings are. Some experts call it “human-level intelligence” rather than AGI, but overall, the concept is the same.
The impact of AGI on humanity may not be as clear-cut as other technology developments, but by demystifying it, we can more honestly evaluate the outcomes. We must question how we get there and what it might look like at the end of this journey.
It’s entirely within our control to embed methods to explain the outputs, control how AGI works, and ensure it benefits humanity.
Why should you care?
You might understandably see AGI as science fiction, as you can’t relate to the day-to-day applications. Let’s look at what’s available and delivering value today. AI tools based on machine learning and computer vision support many aspects of our work today, and Generative AI is already assisting us with applications at work.
For example, we shared at Oracle Cloudworld 2023 how it’s possible to provide a request for information (RFI) to a generative AI chatbot and generate a schedule in Oracle Primavera Cloud. Magic! We also have been working on automating data capture and delivering workflows to the Oracle Construction Intelligence Cloud Service for years. From safety inspections to site progress, machine learning and computer vision are helping with many aspects of project delivery today, and those processes will only continue to advance.
One day, when we achieve AGI or human-level intelligence, these tools will do more than assist our work. Here's a specific example. For testing at the Oracle Industry Lab, we developed a use case where an autonomous drone can fly, take photos, and save the data to the cloud. In this use case, we installed a decal (image) of rust on a steel column to simulate damage. Using AI, the drone was able to detect rust on the column and register that data to applications so the humans in charge can intervene. This work allowed us to improve the process at the Lab to deliver a repeatable autonomous solution. In this scenario, AI couldn’t understand that rust is unnatural (an image on the column) as the technology doesn’t contextualize information, and data is fed from one source (in this case, a photo). The key here is AI today cannot understand our world, but it can process images, videos, and other data based on our model training.
Imagine if AGI can perform the same process. However, in its case it would also be able to understand our world, make decisions to prioritize necessary work based on other data points, explain its findings, plan specific tasks, and ensure that they are delivered to meet the objectives like an expert would. Once all that is possible, that will be the day machines will start replacing us in the workplace, and AI will be the expert.
Looking forward
The projected timeline to achieve AGI ranges from five to 50 years, depending on the expert. The impact of machine learning and computer vision is still considerable even today. Soon, we will have more applications leveraging the power of generative AI.
In a few years, we will see more and more humanoid robots working alongside us. They won’t have AGI, but they will perform high-risk, repetitive tasks alongside us with AI. Humanoid robots' impact on our industry will be significant because, finally, robots will move around our world and work with us.
We built our world for humans—doorways, ceilings, seats, walkways, shelves, locks, etc. Our worksites also follow the same principles: access points, hoists, storage, etc. that are all designed for humans. Humanoid robots with a similar form factor to humans will be able to work alongside us without modifying the physical environment.
Not a week goes by when we don’t hear about a new announcement from a humanoid robotic company. Machine learning, computer vision, and reinforcement learning combined with advancements in mechanical systems, edge and cloud, connectivity, sensors, and batteries are getting us closer to achieving humanoid robots. Soon, they will be untethered and free to work with us.
In the foreseeable future, we will have massive productivity and safety gains by leveraging new AI tools and robotics on our sites. For the built environment professionals, I would recommend focusing on learning and adopting these AI tools. Because until we achieve AGI, AI won't replace our jobs, but someone using AI might.
Burcin Kaplanoglu is Head of Innovation at the Oracle Industry Lab in Deerfield, Ill. He is active in industry organizations and on LinkedIn where he provides educational content related to technology, innovation, robotics, AI and industry use cases.
