Incident: Robots Vulnerable to Hacking: Safety Risks and Privacy Concerns

Published Date: 2017-08-22

Postmortem Analysis
Timeline 1. The software failure incident involving hacker attacks on robots was reported in the article published on August 22, 2017 [62201]. Therefore, the software failure incident happened in August 2017.
System 1. Universal Robots' "collaborative" robots - Multi-jointed arms without real authentication and easily-cracked integrity checks [62201] 2. Alpha2 robot by UBTech - Running a version of Google's Android operating system without code-signing and encryption [62201] 3. NAO robot by Softbank - Vulnerabilities due to premature release of development version code [62201]
Responsible Organization 1. Hackers Lucas Apa and Cesar Cerrudo were responsible for causing the software failure incident by demonstrating hacker attacks against popular robots at the Hack in the Box security conference [62201].
Impacted Organization 1. Universal Robots' "collaborative" robots were impacted by the software failure incident [62201].
Software Causes 1. Lack of real authentication and easily-cracked integrity checks in the software of Universal Robots' collaborative robots, allowing unauthorized access and manipulation of critical safety settings [62201]. 2. Absence of code-signing and encryption in the software of the Alpha2 and NAO companion robots, enabling the installation of rogue software and man-in-the-middle attacks [62201].
Non-software Causes 1. Lack of real authentication and easily-cracked integrity checks in the robots' software [62201] 2. Vulnerabilities in the design and implementation of the robots' operating systems, such as the lack of encryption and code-signing measures [62201] 3. Premature release of development versions of the robots' code by the manufacturers [62201]
Impacts 1. The software failure incident allowed hackers to gain unauthorized access to the operating system of Universal Robots' collaborative robots, enabling them to overwrite critical safety settings, potentially causing the robots to harm human workers within reach [62201]. 2. The incident exposed vulnerabilities in the security measures of companion robots like Alpha2 and NAO, allowing hackers to fully control these robots, manipulate their cameras and microphones, and invade users' privacy [62201]. 3. The researchers' hacking demonstrations revealed over 50 hackable security vulnerabilities in various robots and robotics software, indicating a widespread issue in the industry [62201].
Preventions 1. Implementing strong authentication measures and encryption protocols in the software of the robots could have prevented unauthorized access and control by hackers [62201]. 2. Regularly updating the software of the robots to patch security vulnerabilities could have prevented potential exploits by hackers [62201]. 3. Conducting thorough security assessments and penetration testing on the robots' software before market release could have identified and addressed vulnerabilities proactively [62201].
Fixes 1. Implementing real authentication and robust integrity checks in the software of the robots to prevent unauthorized access and malicious updates [62201]. 2. Utilizing code-signing and encryption measures in the operating systems of the robots to enhance security and prevent the installation of rogue software [62201]. 3. Regularly updating the firmware of the robots to patch security vulnerabilities and enhance overall cybersecurity [62201].
References 1. Security researchers Lucas Apa and Cesar Cerrudo at IOActive [62201]

Software Taxonomy of Faults

Category Option Rationale
Recurring one_organization, multiple_organization (a) The software failure incident having happened again at one_organization: - The article mentions that the researchers Lucas Apa and Cesar Cerrudo demonstrated hacker attacks against robots, including the Alpha2 and NAO robots, which are products of UBTech and Softbank respectively [62201]. - The researchers found vulnerabilities in the robots' software, indicating that similar incidents have occurred with these companies' products before. (b) The software failure incident having happened again at multiple_organization: - The article states that the researchers found more than 50 hackable security vulnerabilities in robots and robotics software sold by companies including UBTech, Softbank, Rethink Robots, Robotis, and Arsatec [62201]. - Additionally, another team of researchers from Italy's Politecnico Milano showed vulnerabilities in the ABB IRB140 industrial robot arm, indicating that similar incidents have occurred with other companies' products as well.
Phase (Design/Operation) design, operation (a) The software failure incident related to the design phase is evident in the article. The security researchers demonstrated hacker attacks they developed against popular robots, including domestic robots and an industrial robotic arm. They found vulnerabilities in the robots' software design, such as lack of real authentication, easily-cracked integrity checks, and absence of encryption in connections. For example, they were able to gain unauthorized access to the robot arm's operating system through a common security vulnerability called a "buffer overflow" and overwrite critical safety settings, potentially causing physical harm to humans working alongside the robot [62201]. (b) The software failure incident related to the operation phase is also highlighted in the article. The security researchers were able to fully control the smaller companion robots by installing software on them. They demonstrated that the robots could be manipulated to send audio and video to a remote spy, turning them into surveillance devices. Additionally, they showed how they could hijack a robot's commands to see through its cameras and control their direction, highlighting the privacy invasion that could occur due to operational vulnerabilities in the robots [62201].
Boundary (Internal/External) within_system, outside_system (a) within_system: The software failure incident discussed in the articles is primarily within the system. The security researchers demonstrated hacker attacks against various robots, including domestic and industrial robots, by exploiting vulnerabilities within the robots' software. For example, they were able to hack the robots to change critical safety settings, send unauthorized commands, and even control the robots fully. The vulnerabilities found in the robots' software, such as lack of authentication, easily-cracked integrity checks, and absence of code-signing, allowed the researchers to manipulate the robots' behavior [62201]. These internal software weaknesses within the robots themselves led to the potential for serious consequences, including physical harm to humans and privacy invasion. (b) outside_system: The software failure incident is also influenced by factors outside the system. The articles highlight the role of human hackers in exploiting the vulnerabilities present in the robots' software. The threat posed by hackers who can take control of robots, manipulate their actions, and potentially cause harm to humans is a significant external factor contributing to the software failure incident [62201]. The actions of external entities, such as malicious hackers, play a crucial role in triggering the software failures in this case.
Nature (Human/Non-human) non-human_actions, human_actions (a) The software failure incident occurring due to non-human actions: The article discusses how security researchers were able to demonstrate hacker attacks against popular robots, such as the Alpha2, NAO, and a robotic arm sold by Universal Robots. The vulnerabilities in the robots' software, such as lack of real authentication and easily-cracked integrity checks, allowed the researchers to gain unauthorized access and control over the robots [62201]. (b) The software failure incident occurring due to human actions: The same article highlights how the vulnerabilities in the robots' software were exploited by human hackers. The security researchers were able to hack into the robots, change critical safety settings, send them unauthorized commands, and even turn them into surveillance devices that could transmit audio and video to a remote location. These actions were all initiated by human hackers, showcasing the potential dangers of such exploits [62201].
Dimension (Hardware/Software) hardware, software (a) The software failure incident occurring due to hardware: - The article discusses a software failure incident related to robots, specifically the Universal Robots' "collaborative" robots, where the researchers found that the robots' software had no real authentication and only easily-cracked integrity checks meant to prevent a hacker from installing malicious updates. They were able to gain unauthorized access to the robot arm's operating system through a common security vulnerability called a "buffer overflow" and overwrite critical safety settings, potentially causing physical harm to both the robot itself and human workers within reach [62201]. (b) The software failure incident occurring due to software: - The same article also highlights software failures in smaller companion robots like the Alpha2 and NAO, where the researchers were able to fully control these robots by installing software on them. Vulnerabilities such as lack of code-signing in the Alpha2's Android operating system and encryption issues in the NAO robot's code allowed for malicious apps to be injected and for control of the robots' cameras and microphones, leading to privacy invasion concerns [62201].
Objective (Malicious/Non-malicious) malicious (a) The software failure incident described in the articles is malicious in nature. The incident involves security researchers demonstrating hacker attacks against popular robots, such as the Alpha2, NAO, and a robotic arm sold by Universal Robots, with the intent to show how these machines can be hacked to change critical safety settings, turn into surveillance devices, and potentially cause physical harm to humans [62201]. The researchers were able to exploit vulnerabilities in the robots' software, such as lack of authentication, easily-cracked integrity checks, and absence of security measures like code-signing and encryption, to gain unauthorized control over the robots [62201]. The potential consequences of these malicious attacks include catastrophic harm to both the robots themselves and human workers in industrial settings [62201]. (b) The software failure incident is non-malicious in the sense that the vulnerabilities exploited by the security researchers were not intentionally introduced by the manufacturers of the robots. The researchers identified over 50 hackable security vulnerabilities in robots and robotics software from various companies, including UBTech and Softbank, and initially obscured the specific vulnerabilities to give the manufacturers a chance to address them [62201]. However, the researchers found that the security issues persisted even after monitoring updates for the robots and did not necessarily depend on users setting strong passwords or securing their Wi-Fi networks [62201]. This indicates that the vulnerabilities were not deliberately introduced by the manufacturers but were present due to oversight or inadequate security measures.
Intent (Poor/Accidental Decisions) poor_decisions (a) The intent of the software failure incident related to poor decisions is evident in the article. The security researchers demonstrated hacker attacks against popular robots, including domestic and industrial robots, by exploiting vulnerabilities in their software. For example, they found that the robots' software had inadequate authentication and easily-cracked integrity checks, making them susceptible to unauthorized access and control [62201]. These poor decisions in the design and implementation of the robots' software contributed to the potential risks posed by the hacking attacks.
Capability (Incompetence/Accidental) development_incompetence, accidental (a) The software failure incident related to development incompetence is evident in the article where security researchers Lucas Apa and Cesar Cerrudo demonstrated hacker attacks against popular robots, including domestic and industrial robots. They found vulnerabilities in the robots' software, such as lack of real authentication and easily-cracked integrity checks, which allowed them to gain unauthorized access and control over the robots. For example, they were able to overwrite critical safety settings on the industrial robot arm, potentially causing harm to human workers [62201]. (b) The software failure incident related to accidental factors is highlighted in the article where the researchers discovered vulnerabilities in the robots' software that were not intentionally designed but existed due to oversight or premature release of the products. For instance, the Alpha2 robot by UBTech ran a version of Google's Android operating system that lacked code-signing, making it susceptible to rogue software installation. Similarly, the NAO robot by Softbank had vulnerabilities in its code that were not adequately addressed before being pushed to market [62201].
Duration permanent (a) The software failure incident described in the articles is more aligned with a permanent failure. The security researchers demonstrated hacker attacks against various robots, showing how they could gain unauthorized access, change critical safety settings, and fully control the robots [62201]. These vulnerabilities in the robots' software and lack of proper authentication mechanisms make them susceptible to being hacked, potentially leading to catastrophic consequences such as causing harm to humans or invading privacy. The researchers highlighted that the security issues observed in the robots were not adequately addressed by the manufacturers, indicating a persistent and ongoing risk of exploitation [62201].
Behaviour crash, omission, value, other (a) crash: The article describes a scenario where the researchers were able to gain unauthorized access to a robot arm's operating system and overwrite critical safety settings, potentially causing the robot to damage itself or harm human workers within reach. This could lead to a crash scenario where the system loses its state and fails to perform its intended functions, endangering both the robot and humans [62201]. (b) omission: The researchers demonstrated that they could fully control smaller companion robots meant for entertainment and education by installing software on them. They were able to manipulate the robots' cameras and microphones, intercepting and moving data around a target's house. This indicates an omission failure where the system omits to perform its intended functions of protecting user privacy and security [62201]. (c) timing: The articles do not provide specific information about a timing failure where the system performs its intended functions but at the wrong time. (d) value: The researchers found vulnerabilities in the software of the robots that allowed them to install rogue software, inject malicious apps, and gain unauthorized control over the robots. These actions indicate a value failure where the system performs its intended functions incorrectly by allowing unauthorized access and control [62201]. (e) byzantine: The articles do not mention a byzantine failure where the system behaves erroneously with inconsistent responses and interactions. (f) other: The behavior of the software failure incident can also be categorized as a security vulnerability. The vulnerabilities found in the robots' software allowed the researchers to hack into the robots, change critical safety settings, intercept data, and gain unauthorized control. These security weaknesses pose a significant risk to both the robots and the users, highlighting a critical aspect of the software failure incident [62201].

IoT System Layer

Layer Option Rationale
Perception sensor, actuator, processing_unit, network_communication, embedded_software (a) sensor: The software failure incident discussed in the articles is related to the sensor layer of the cyber physical system. The security researchers demonstrated attacks against robots by manipulating their sensors, such as infrared sensors detecting someone nearby and cameras and microphones being used for surveillance purposes [62201]. (b) actuator: The failure was also related to the actuator layer of the cyber physical system. The researchers were able to overwrite critical safety settings of the industrial robot arm, affecting its movements, speed, force, and reactions, potentially causing harm to both the robot itself and human workers within reach [62201]. (c) processing_unit: The incident involved failures related to the processing unit of the cyber physical system. The security researchers exploited vulnerabilities in the robots' software, including using a common security vulnerability called a "buffer overflow" to gain unauthorized access to the robot arm's operating system and manipulate critical files that control the robot's behavior [62201]. (d) network_communication: The failure also encompassed issues related to network communication within the cyber physical system. The researchers demonstrated man-in-the-middle attacks on the robots, intercepting and manipulating data from the robots' cameras and microphones, showcasing the potential risks of privacy invasion through compromised network communication [62201]. (e) embedded_software: The software failure incident was significantly influenced by errors in the embedded software layer of the cyber physical system. The security researchers found vulnerabilities in the robots' software, such as lack of authentication, easily-cracked integrity checks, absence of code-signing, and encryption, which allowed them to install malicious software and fully control the robots [62201].
Communication connectivity_level The software failure incident discussed in the articles is related to the communication layer of the cyber physical system that failed at the connectivity_level. The incident involved vulnerabilities in the communication protocols and security measures of various robots, allowing hackers to take control of the robots, manipulate their movements, and access their cameras and microphones remotely. The vulnerabilities included lack of authentication, easily-cracked integrity checks, absence of code-signing, and unencrypted connections, which are all related to network and transport layer security issues [62201].
Application TRUE The software failure incident described in the articles is related to the application layer of the cyber physical system. The failure was caused by vulnerabilities in the robots' software, including lack of authentication, easily-cracked integrity checks, and security vulnerabilities such as buffer overflow. These vulnerabilities allowed hackers to gain unauthorized access to the robots' operating systems, change critical safety settings, and fully control the robots, potentially causing physical harm to humans and invading privacy [62201].

Other Details

Category Option Rationale
Consequence harm, property, non-human, theoretical_consequence, other (a) death: The software failure incident discussed in the articles did not result in any reported deaths. (b) harm: The software failure incident had the potential to cause physical harm to humans. The researchers demonstrated how hacking into robots, such as the Universal Robots' collaborative robots, could lead to catastrophic consequences, including causing bone fractures and harming human workers within reach [62201]. (c) basic: The software failure incident did not impact people's access to food or shelter. (d) property: The software failure incident had the potential to impact people's material goods, particularly in industrial settings where robots are used alongside humans. For example, hacking into the Universal Robots' robotic arm could lead to damaging the robot itself and potentially causing harm to human workers [62201]. (e) delay: The software failure incident did not result in any reported delays in activities. (f) non-human: The software failure incident directly impacted non-human entities, specifically robots. The researchers demonstrated how they could hack into robots, such as the Alpha2 and NAO, to control them, turn them into surveillance devices, and manipulate their cameras and microphones [62201]. (g) no_consequence: The software failure incident had real observed consequences, particularly in terms of the potential physical harm that could result from hacking into robots. (h) theoretical_consequence: The articles discussed potential consequences of the software failure incident, such as the serious physical harm that could occur if robots were hacked, but there were no reported instances of these theoretical consequences actually happening [62201]. (i) other: The software failure incident also raised concerns about privacy invasion, as hackers could intercept, manipulate, and move data from the robots' cameras and microphones, presenting a realistic worry beyond just physical harm [62201].
Domain information, other (a) The failed system was related to the information industry as it involved security researchers demonstrating hacker attacks against robots used for domestic and industrial purposes [62201]. (b) The failed system did not directly relate to the transportation industry. (c) The failed system did not directly relate to the natural resources industry. (d) The failed system did not directly relate to the sales industry. (e) The failed system did not directly relate to the construction industry. (f) The failed system did not directly relate to the manufacturing industry. (g) The failed system did not directly relate to the utilities industry. (h) The failed system did not directly relate to the finance industry. (i) The failed system did not directly relate to the knowledge industry. (j) The failed system did not directly relate to the health industry. (k) The failed system did not directly relate to the entertainment industry. (l) The failed system did not directly relate to the government industry. (m) The failed system was related to the robotics industry, which falls under the "other" category as it is not explicitly mentioned in the provided options [62201].

Sources

Back to List