Why Contractors Need to Become Influences

The next phase of the Artificial Intelligence story is now underway and will require companies to take action today if they want to have an advantage over their competitors tomorrow. 

AI is clearly starting to move into the physical world. Waymo and Tesla are expanding their autonomous cars into more and more cities, OpenAI struck a deal to buy IO, a company from legendary Apple hardware designer Jony Ive, and Meta is opening up retail stores. 

But how do companies take advantage of this shift even as it is not fully developed? The most obvious opportunity is the difference in training data from LLMs. Humans have written potentially billions of different pieces of text since the invention of the printing press. With such a large training data set, it is no wonder that LLM chatbots were one of the first manifestations of AI. However, as AI moves into the physical world, AI starts to interact through video rather than text. Although it may seem like there is an endless amount of videos on Youtube to train AI on, how many of these videos are shot from the first person? An even better question, how many of these videos are shot from the first person in which a person performs a specific task? 

This is where the first mover opportunity rests. Companies are able to start to capture their own first person video that can later be used as a company specific AI training data set. As far-fetched as using first person video to train AI sounds, Tesla has already accomplished this. Click on the video below and see what we mean. 

So what sort of future applications or products could be developed that would be useful for construction firms that would justify the effort of a data collection program?

Construction workers are already mandated by OSHA to wear certain personal protective gear while on the job site, such as hard hats or glasses. Meta and Anduril have recently teamed up to create AI enabled, virtual reality helmets and glasses for the U.S. Militarily. How long until these products are adopted for civilian use? Given that construction is one industry that requires the use of these items, it just makes sense that these technologies will eventually find their way into the industry.

Meta and Andruil are developing wearables that increase soldiers sight and hearing, which may have direct uses for construction, however what are potential industry specific applications?

One potential application is using AI for quicker equipment repairs and less reliance on specialized knowledge. Equipment breakdowns cost contractors valuable time and time that is not spent on productive work is time that costs contractors money. What if there were a pair of glasses that would be able to walk a new hire through an equipment repair? The glasses would have a camera and speaker and the AI model would be trained on not only equipment manuals but thousands of hours of first person video of similar repairs. 

The idea would be that the wearer would describe the issue with the machine with words and also video demonstrations. The AI would examine the video feed and the voice commands, then cross reference these inputs with the stored equipment manuals and the library of previous repairs. From there, the AI would use voice responses to guide the mechanic where to go on the machine and what to do. The AI model would be able to recognize hand placements and guide the user to the correct spot on the machine with voice commands.

This is just one application that would reduce training costs, reduce downtime costs and also increase safety. But there are plenty of others from AI guided installation or equipment use.

 The key really is the data that would be available for the AI to train on. Companies that act now will be able to leverage AI to generate far greater returns than their competitors. Those that resist the AI transformation risk being overtaken.