An infectious disease is any disease caused by a pathogen, such as a virus, bacteria, a parasite, or a fungus. Currently, the world is battling the Covid-19 pandemic, which has caused more than 7 million cases of infection and close to 200,000 deaths in the US alone as of September 2020. Notable infectious disease epidemics in the recent past before Covid-19 include SARS (2003) and Ebola (2013–2016), although they were less severe.‘Infectious disease technology’ refers to 1) camera-based detection tools that use thermal imaging and facial recognition, 2) surveillance gadgets (smart wearables), and 3) software platforms that convert data from hardware (described in points one and two above) to identify individuals (i.e. to trace contacts and share system-generated insights in real-time). The software tools often include the latest artificial intelligence (AI) and machine learning capabilities that allow end-to-end, seamless, actionable insights.(Note: This report does not cover clinical/medicinal diagnostics methods, such as PCR testing, as they are unlikely to be used outside of the healthcare industry).

These detection tools are now being promoted to the masses in response to the Covid-19 pandemic, including for implementation in office spaces. The smart wearables segment has an innovative set of products (smart badges, bands) for containing Covid-19. Players often provide the hardware and software parts in combination, although several operate as pure-play software providers. (e.g., Landing.ai). Development and integration of AI and machine learning act as main enabling factors

AI and machine learning have been used in the surveillance industry for most of the past decade. Interest was reignited recently by the Covid-19 pandemic, for which faster and accurate disease detection tools are needed. For instance, unlike traditional thermal cameras, thermal fever cameras with built-in AI and machine learning can screen people for body temperature more accurately (roughly ±0.3°?/roughly ±0.6° F), identify potential risk and threat targets, and send that data to stakeholders in real time.AI and machine learning have been enabled largely by other technological advancements such as the following:

Advanced chips for parallel computation: Until early this decade, computers that ran AI software could only run one process at a time. This changed with the creation of a chip called the ‘graphic processing unit’ (GPU) which would later be included in PC motherboards. Made for visual processing, it can recalculate millions of pixels multiple times in a second. In 2009, the chip was used for running neural networks parallelly, aiding AI development, as hundreds of millions of connections could be made between different data structures (nodes). Neural networks are the backbone of any AI technology platform. Those running on GPUs are routinely used by cloud-enabled companies to develop AI software for various applications. This has caused an increase in the production of GPUs, resulting in lower chip prices. Examples of key GPU chips include Alphabet’s tensor processing unit (TPU), Qualcomm’s neural processing unit (NPU), Nvidia’s deep learning chip, and IBM’s TrueNorth neuromorphic computing platform, used in the development of AI in thermal scanners, among other machines.

Big data: The concept of AI and deep learning centers on datasets used for categorizing variables. Data collection plays a vital role in the development (and in teaching) of AI in databases, web cookies, online footprints, storage, and search results. For instance, AI software in a thermal scanner might find changes in a scanned individual's behavior, or any behavior that raises suspicion (e.g., excessive sweating, etc.), based on its learning from datasets available on the world wide web.

Improved algorithms: AI and machine learning make use of algorithms—relationship combinations stacked in layers. Algorithms have improved dramatically over the years as data scientists used complex mathematical inputs that helped learning accumulate faster, proceeding up the stack of layers. Deep learning and AI algorithms have developed even more rapidly when integrated with GPUs or AI software chips.Nevertheless, the development of chips, big data, and algorithms have also improved over the past decade as long-term evolution (LTE) technology evolved from 3G to 5G, enabling faster data processing.