Chris Bishop, Marketing Director at Ipsotek calls for the increased adoption of AI-powered video analytics.
Despite the significant challenges presented by COVID-19, nearly half of all artificial intelligence (AI) investments have remained unaffected since the start of the pandemic. In fact, just under a third (30%) of organisations are actually planning to increase investment in AI, according to recent research from Gartner.
Additionally, the latest Emerging Tech Executive Report from RELX, which analysed the adoption of AI during the pandemic, found that 68% of businesses increased AI spend during this period, with nearly half (48%) investing in new AI technologies. Which is not surprising given that the pandemic has shown the potential of AI for many businesses. Applications for AI include automating workflows, enhancing customer experience, predictive analytics and in security fraud detection. However, during the pandemic, its benefits really came to the fore, with businesses adopting chatbots to optimise communications with customers; deploying video analytics to ensure social distancing is maintained within workplaces and adhered to, with the added benefit of being able to use Forensics tools for contact tracing.
All this has led to the UK government providing a £20 million funding boost to initiatives and research that will keep the UK at the forefront of this technological innovation. Beyond this however, the demand for AI is also growing across smart cities, according to a new report released by global tech market advisory firm ABI Research. In its latest Deep Learning-Based Machine Vision in Smart Cities application analysis report, ABI Research estimated that the global installed base of smart cameras with an AI chipset is set to reach over 350 million by 2025.
Looking into the research, nearly two-thirds of these cameras are predicted to feature a minimum of one AI chipset, with this surge attributed to increasing demand from global governments to integrate AI into smart city development. Such utilisation is anticipated to improve and automate decision-making across many applications, including smart traffic management, perimeter security and for monitoring pedestrian flow to name a few. Several US banks have even started to deploy AI to monitor customers and workers.
Barriers to adoption
Despite this, public trust and privacy concerns remain a barrier to widespread adoption, particularly around technologies such as face recognition. Here in the UK, during the last 18 months, we have witnessed a significant backlash amongst the general public after reports emerged in the national media that images of several individuals obtained via a face recognition system at King’s Cross Station in London were shared by local police without the prior knowledge of the Metropolitan police or the Mayor’s office.
These concerns are not restricted to the UK, with public perception in the US – where the facial recognition industry is set to grow to reach US$7bn by 2024 according to market research firm Markets and Markets – it is also a major issue. The use of such technology for surveillance in particular is a major topic of contention, with an absence of federal regulations causing public concern about the accuracy of the technology and any biases in its usage by law enforcement agencies.
In line with this, various US States have implemented different legislation regarding the use of the technology, including banning it altogether in some places such as in San Francisco, California. Legal proceedings gathered further momentum in the US at the start of 2021, with officials attempting to reinstate a previously unsuccessful Bill to prevent federal agencies from using face recognition software.
Elsewhere in the world, such as in Singapore for example, the use of facial recognition is becoming ever growingly prevalent. Indeed, in November 2020 the government announced that facial recognition along with biometric scanning procedures would be introduced to all of the country’s immigration checkpoints as part of Singapore’s overall approach to border security.
A recent survey carried out by the journal Nature meanwhile, sought to obtain the view of academics when it comes to privacy concerns and ethics surrounding facial recognition. Nearly 500 researchers who had previously published papers on the technology or the wider fields of AI and computer science contributed, with approximately two thirds saying that the use of the technology to determine or predict personal characteristics (ethnicity, age and gender etc) should only take place with the prior consent of the individuals involved.
Feeling the benefit
While all of these concerns and reservations need to be heard and taken into consideration, we must not also lose sight of the fact that there are a myriad of technologies available today that can deliver meaningful insights without putting an individual’s privacy at risk.
For example, research published in 2020 revealed that London has 627,727 cameras for 9.3 million residents – the equivalent of 67.5 cameras per 1,000 people. The data obtained and stored by these cameras will have many different uses and be extremely valuable, but with such a volume of footage it can be very hard to harness.
In these contexts, solutions such as A.I.V.A. (Artificial Intelligence Video Analytics) can provide the answer in terms of detecting patterns of behaviour. In the last 12 months during the COVID-19 pandemic, this includes use cases such as identifying breaches of social distancing so that steps can be taken to ensure this doesn’t become commonplace.
A.I.V.A. solutions utilise existing camera networks and geospatial algorithms to determine an individual’s location in the camera field of view in real-time by learning the perspective of the scene and calculating the GPS coordinates of the individual based on their location in the camera’s field of view. Most critically however, these calculations are made without reliance on identifying any specific individual(s) thorough face recognition and nor do they require access to a mobile phone signal or compromise any personal data. Essentially, the technology serves as a solution for all those concerned with privacy.
In the context of usage for identifying social distancing breaches for example, adopting A.I.V.A. solutions can enable a business, organisation, or in the case of ABI’s Research, a government smart city project, to establish a set of pre-defined criteria in order to detect when one or more persons are in breach of guidelines. These incidences can then be automatically logged, trigger an alert and facilitate an action (e.g. an alarm or audio warning) without recording personal details or revealing the identities of the individuals involved.
Admittedly, this might not be particularly impactful in a single, isolated occurrence, but generating a report or multiple reports built upon a series of instances gathered from multiple hours of CCTV footage across thousands of cameras in a network could provide many beneficial insights.
Smart city applications
In the case of smart cities and monitoring pedestrian flow specifically, this might include the identification of areas where people are frequently congregating, the most commonly used exit of a building or train station, or locations attracting large queues at a particular time of the day. In the era of COVID-19, the benefits of this are clear, including the capability to conduct rapid contact tracing in order to identify individuals who may have come into contact with someone who later has been identified as testing positive for COVID-19.
In many countries the task of searching through records and CCTV footage to identify who someone may have come into contact with is labour intensive and prone to error. By combining the output of a social distancing applications with A.I.V.A., stored video clips can be rapidly searched against specific criteria such as date, time, location, colour of clothing, etc with the results produced in seconds. If this approach were then combined with the mobile applications now available in many countries, contact tracing accuracy would undoubtedly improve.
It’s clear there is a huge untapped potential in the adoption of A.I.V.A. solutions, as businesses and government organisations alike welcome the operational value AI can provide for mission critical applications and everyday activities. Solutions that don’t compromise an individual’s privacy are best placed to support the development of smart cities around the world.
This article was originally published in the May edition of Security Journal UK. To read a free digital copy of the magazine, click here.