Skip to Main Content

Strategic Business Insights (SBI) logo

Internet of Things November 2020 Viewpoints

Technology Analyst: David Strachan-Olson

Cyber-Physical Vulnerabilities

Why is this topic significant?

The May 2020 Viewpoints discusses how key forces in the global business environment affect the commercialization of opportunities related to the Internet of Things. This Viewpoints provides a concise analysis of one such force—cyber-physical vulnerabilities—and describes plausible alternative pathways along which this force could develop.

Description

Most current-generation cyberattacks' direct impacts are limited to the digital realm, as is the case with theft of data from a database or corporate intranet. However, many emerging digital systems—including driverless vehicles, robots, automated factories, and smart grids—take action in the physical world through actuators, motors, switches, and other types of components. Malicious actors can exploit these cyberkinetic components to cause real physical damage to objects and people.

Prominent examples of cyber-physical attacks exist, such as the Stuxnet computer worm attack on Iranian centrifuges, and exploits of General Motors' OnStar in-vehicle system that could enable a hacker to bring a vehicle to a stop. Additionally, ransomware attacks, which aim to lock out system operators, against critical infrastructure and systems are becoming more common. Although ransomware attacks do not necessarily cause physical damage, the shutdown of a factory or warehouse because of an infected computer system can have physical knock-on impacts. Additionally, ransomware attacks can infect—and have infected—health-care systems and equipment, which can lead to patient death. In September, authorities in Germany launched a "negligent homicide" investigation after a patient died as a result of a ransomware attack at a hospital.

Implications

As organizations use more connected systems because of their advanced capabilities, cybersecurity issues will become more apparent. Cybersecurity is not a single battle but an endless war. Operating systems, drivers, software, and web interfaces are changing continuously to add new features, but these changes also raise new security issues. Additionally, bad actors continue to discover new vulnerabilities in hardware and software enabling new ways to invade systems. Companies deploying critical systems will likely employ moderate cybersecurity practices to ward off automated attacks but will remain vulnerable to attackers with sufficient resources. Potentially, new technologies and open industry standards could address many common vulnerabilities and make most systems sufficiently safe. Alternatively, sustained attacks against cyber-physical systems could change the direction of technological development and encourage companies to take a cybersecurity-first approach. High-profile incidents could also spur increased government regulation and increase public skepticism about cyber-physical systems.

Impacts/Disruptions

Many stakeholders envision a significant increase in the number of cyber-physical systems as innovations in Internet of Things, Industry 4.0, 5G, autonomous vehicles, and robotics continue to occur. However, the threat of cyber-physical attacks could force companies to take cybersecurity more seriously, which would improve safety but could limit the functionality of systems. For example, autonomous-vehicle makers might decide remote operation is too big a cybersecurity risk and never develop the feature. In extreme cases, companies might seek to develop self-contained, air-gapped systems and private communication networks.

Scale of Impact

  • Low
  • Medium
  • High
The scale of impact for this topic is: Medium to High

Time of Impact

  • Now
  • 5 Years
  • 10 Years
  • 15 Years
The time of impact for this topic is: Now to 10 Years

Opportunities in the following industry areas:

IoT device supply, cybersecurity, network security, IT, enterprise IoT, smart cities, connected infrastructures, connected homes, compliance and verification

Relevant to the following Explorer Technology Areas:

Applicability of Computer Vision

Why is this topic significant?

The May 2020 Viewpoints discusses how key forces in the global business environment affect the commercialization of opportunities related to the Internet of Things. This Viewpoints provides a concise analysis of one such force—applicability of computer vision—and describes plausible alternative pathways along which this force could develop.

Description

Many stakeholders attribute the current resurgence in artificial-intelligence research to a 2012 paper by Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton that described image classification using an implementation of deep convolutional neural networks on graphics-processing units. Since then, image recognition and computer vision have advanced significantly and continue to represent one of the fastest-advancing areas of modern AI research. Image recognition and computer vision have a seemingly infinite number of potential applications in areas ranging from medical imaging to autonomous vehicle sensing to facial recognition.

In some cases, AI in the cloud carries out complex computer-vision tasks. Cloud-service providers have raced to introduce intelligent camera solutions, which include Amazon Web Service's DeepLens and Microsoft's Azure Kinect. However, continuing improvements to AI models and dedicated AI hardware are enabling developers to implement on-device computer-vision features. Simple computer-vision models are already appearing in a range of devices, including smartphones, security cameras, and robots. Dedicated vision processors, such as those from Qualcomm and Intel Movidius, are likely to enable the proliferation of computer vision to many more applications.

Implications

The continuing advancement of computer-vision algorithms and the availability of low-cost vision processors could lead to the adoption of intelligent cameras across a wide range of application areas including autonomous vehicles, robotics, connected homes, Industry 4.0, and smart cities. The functionality of intelligent cameras could continue to evolve with the addition of depth sensors, hyperspectral imaging (for analyzing nonvisible light spectrums), event cameras, and light-field cameras. Alternatively, computer-vision systems may hit a technological bottleneck that makes ensuring that systems are robust and free of bias nearly impossible for engineers. Computer-vision would continue to have uses beyond those in critical applications, however.

Impacts/Disruptions

Advanced computer-vision systems could collect vast amounts of information about the environment and the objects and people in it. Many stakeholders' visions for the Internet of Things (IoT) involve selling hundreds of millions of electrical components—including microprocessors, radio-frequency equipment, batteries, and sensors—to connect many everyday objects. However, in many sensing applications, a network of intelligent general-purpose cameras could monitor and sense the state of the world and many objects without the need to connect every object. This capability would represent a dramatic shift from the traditional concept of the IoT.

Scale of Impact

  • Low
  • Medium
  • High
The scale of impact for this topic is: High

Time of Impact

  • Now
  • 5 Years
  • 10 Years
  • 15 Years
The time of impact for this topic is: Now to 10 Years

Opportunities in the following industry areas:

Computer vision, AI processors, cloud services, surveillance systems, smart cities, retail, manufacturing, warehouses, robots, personal assistants

Relevant to the following Explorer Technology Areas: