Jobs of the future: how self-piloted AI drones are creating exciting new opportunities
Not too long ago, automated flying machines whizzing around the skies were concepts gracing the pages of science fiction novels.
Today, drones are used and seen frequently. From capturing incredible aerial shots for documentaries, to inspecting large structures to spot potential weaknesses, drones have become invaluable tools.
As the Wright brothers’ first successful airplane evolved into the sophisticated aircraft we see today, the very first drones too, have evolved and improved – and are continuing to do so, thanks to artificial intelligence.
AI is paving the way for a new generation of self-flying drones, which can carry out tasks without requiring a human operator. The possibilities are endless, ranging from automatic safety inspections, to 24/7 delivery services.
The advent of new technology also provides new opportunities and jobs, many of which do not exist today. We had the pleasure of speaking with Arshia Gratiot, Founder and CEO of Third Space Auto – a company spearheading the AI drone space – to discuss the future of AI-powered drones, and the benefits and potential opportunities they could provide.
The DNA of an AI drone
When most people think of drones, they probably think of an enthusiast piloting a consumer drone to take scenic videos and photos. The reality for more sophisticated use cases however, is far more complex.
“Complex drone operations are not as simple as flying a hobby drone,” Gratiot states. “It’s not unusual to have a team of five people, each with different tasks.” In addition to the pilot, these tasks can also include having someone in charge of planning out the mission, while another maps the topography of the land below. The skies, of course, while free from the clutter of the ground, also have their own congestion, with other aircraft sharing the same space, not to mention a myriad of aviation laws and restrictions.
A true AI-piloted drone must execute all of these roles simultaneously and accurately, while being able to make real-time decisions based on unexpected or changing circumstances. This is a vital component in the work that Third Space Auto is doing.
The company is currently developing AI-powered drones for automated delivery purposes, where users can easily order items for delivery by drone, from an app. A typical journey that can take 45 minutes in a vehicle in a city, can take just five minutes by drone, and the ability for a drone to spot and avoid an unexpected obstacle such as a construction crane, for example, is a paramount requirement for safety and functionality.
Microsoft AirSim – an open-sourced system designed to train autonomous systems – is the tool used to help develop the AI. It allows Third Space Auto to safely simulate realistic environments and drone dynamics in a virtual environment, to help ensure their safe and correct operation in the real world.
The importance of transparency
You’d be forgiven for thinking that Third Space Auto’s AI solution uses deep learning to allow drones to improve and learn on their own, but that’s not the case. “We don’t use deep learning, because it’s a black box, and that’s something we can’t do, because we need to be able to clearly see and trace each decision a drone has made on its journey.”
The black box, in this case, means that in deep learning applications, data goes in, and decisions come out, but the processes in between aren’t visible. Unlike the approach used with machine learning (which involves a system performing a function with data that gets progressively better over time,) deep learning sees algorithms determine if its predictions are accurate, on their own accord.
In the case of AI drones, it’s very important to be able to see each decision clearly. Avoiding an unexpected obstacle and going off-route, for example is something that needs to be documented, to clearly show the reason why a drone has deviated from a pre-approved path.
“We have a machine learning, rule-based system,” Gratiot explains. “It’s ‘if this, then that’ – if a drone is forced to fly higher because of a tree, for example, then we can see exactly what happened, and why.”
As AI drones are developed, new potential opportunities are also beginning to emerge. In Third Space Auto’s case, for example, the drones are, in Gratiot’s words, “…tools at the end of a much larger ecosystem, with a much larger value chain. The drone definitely needs to be applied within an architecture of some form, which needs to be defined and created by a human.” The role for an AI drone architect might not exist in any job listings today, but it’s certainly a likely career opportunity that will develop in future.
Beyond the developing of the AI itself, a world where hundreds of automated delivery drones carrying out their tasks in the city sky could also require people in charge of monitoring their progress and status, in a similar way to the logistics and transport manager roles we see in existing fields such as air traffic control, or public transportation.
A team of specialists monitoring drones could be tasked with ensuring that all units are on safe, regulated flight paths, while running on time and taking the most efficient routes and cargo possible. Some of the skills required for this exist today, but new learnings and increased experience could lead to a whole field dedicated specifically to self-flying drone logistics and management.
The Association for Unmanned Vehicle Systems International (AUVSI) states that “…in the first three years of (commercial unmanned aircraft integration) integration more than 70,000 jobs will be created in the United States with an economic impact of more than $13.6 billion. This benefit will grow through 2025 when we foresee more than 100,000 jobs created and economic impact of $82 billion.”
While this projection is US-focused, it still provides an overview of a future where development in AI-powered aviation could provide many new opportunities for skill development and new employment opportunities.
With the labour market changing and an estimated shortage of 500,000 workers with digital skills in 2020, it’s vital that we embrace innovation such as AI-powered drones, and encourage everyone to explore and embrace these exciting new opportunities.