One of the major challenges facing the global community as it implements AI solutions to address critical challenges will be developing solutions that don’t create a digital divide between those who already have access to technology, and those who don’t. While AI technologies have immense potential to act as a social leveller by bringing down skill barriers, there remains the issue that the technology could exacerbate existing inequalities and fail to improve the lives of the world’s poorest.
Electricity
Access to electricity is a significant challenge in the rural communities of many of the world’s poorest countries. For AI systems to work, you need a steady supply of electricity to power servers and data-gathering devices.
One of the challenges that LUMs found was that the power accessible using the available grid system was providing a consistent supply of electricity to their central servers, but not to their towers distributed throughout the forest. This ran the risk of the hardware on the towers losing power supply at a crucial moment. Their solution combined, solar, sensors, and cloud-based solutions to maintain power throughout the system.
All the equipment on the tower is operated using solar-charged batteries. The tower has a sensor component which contains a camera and weather station. In addition, it has an embedded computer that takes the data from the sensor and transmits it over the network to the cloud storage. All of this is now based on solar-charged batteries. Once the data is transmitted to the cloud it is instantaneously pulled by the Central Server where the AI module detects a smoke or fire event. The cloud, central server, their inter-communication, and the AI module are connected to grid-based power.
Cost
Another challenge to ensuring that AI technologies are leveraged to benefit everyone, and not just those in the best position to take advantage of them, is cost. AI solutions can often rely on expensive hardware, which is unaffordable in the context of many international development programmes. We’ll do a more thorough analysis of the cost of developing computer vision solutions later in the module.
On this pilot, the challenge manifested in the prohibitive cost of thermal cameras. Thermal cameras use sensors to detect infrared energy (heat energy), rather than light, and create a visual heat map. As such they can be incredibly useful for detecting fires in a wide range of contexts, specifically where certain things look like fires but aren’t emitting the same heat energy. However, the cost of the hardware was prohibitive. The team would have had to spend a significant portion of their budget just on the cameras, which as we’ll explore later are only one part of the costing model for computer vision systems.
The team opted instead for PTZ (Pan-Tilt-Zoom) cameras, which capture images through visible light. Gathering images through optical visibility alone does limit functionality. The team resolved this by supplementing the cameras with IoT sensors on the ground to gather more data. The combination of these two lower-cost technologies allowed them to develop a system which worked and was much more cost-effective. Further cost reduction was achieved by having a 30-minute gap between the capturing of images (to reduce the burden of data transmission) which could be adjusted depending on the needs of the situation. These strategies ultimately led to a tenfold reduction in projected cost.
These two strategies demonstrate a general lesson for keeping costs down: expensive, state-of-the-art hardware is not always necessary to achieve the best results. A careful analysis of the technical requirements of specific use cases allows you to identify the hardware needed to deliver good results. Additionally, experimenting by combining multiple cheaper options can be an effective way to deliver results at a lower cost.
In the next section, you’ll learn more about some of the ethical challenges and solutions the team faced.