A number of problems affecting how robots identify commands sent by their real owners has been identified by security researchers.
IOActive Security researchers managed to hack into robots following tests on robotic devices including Softbank’s NAO and Pepper, UBTech’s Alpha and Universal’s UR Robots. Researchers reported finding key vulnerabilities including insecure communications, authentication issues and week cryptography.
Lucas Apa, a security consultant at IOActive said: “We found a lot of authentication vulnerabilities, where the use of the robot is not actually verified to be the owner. Anyone within the same network can interact with the robot and issue commands.”
By simply tracking down a robot’s IP address attackers could hack into the system and take over. After hacking into the network system, researchers at the security firm were able to shut down key safety features which enabled them to take over the robot and turn them into ‘spy devices’.
Many robots are programmed to work alongside humans, to benefit businesses by carrying out everyday tasks. IOActive, however, warned in a report that “Robots in industries could easily become a lethal weapon.”
In severe cases robots could be controlled to find and use weapons that could cause destructive damage and significant harm to humans. IOActive found vulnerabilities across the robot models that would allow hackers access to personal information and other data.
The research comes after Elon Musk joined 116 other tech leaders in an an open letter warning the United Nations technology that AI could lead to ‘killer robots’ being created and urged the UN to take action against the development of dangerous autonomous devices.
“Robots will become more powerful and in the near future we think that robots could do much more harm than they could do today,” said IOActive.