Something that will become more important as domestic robots become ubiquitous, is the responsibility of manufacturers to ensure the databases queried by their robots are as secure as banks. The danger that is possible here is simple and disturbing. The possibility exists for hostile actors to switch the identity of a food ingredient with a toxic household chemical.
It no longer seems like science fiction to imagine a kitchen robot that would prepare a meal or food item, the governing AI will rely upon a hierarchy of understanding most likely housed in the cloud. This is a necessity because when a robot receives instructions to make a cake the variety of what flour can look like through every kitchen in the country requires an understanding of flower that transcends simply reading the table. This transcendent understanding of what flower is will likely be queried from a cloud of the manufacturer where the robot will receive the concept of what it is looking for.
The potential for malicious actors to manipulate this exists in three distinct methods. First the broad approach, where the entire cloud is penetrated and every system is affected. This will be quickly identified and quickly remedied. This is also very similar to what would happen if a malicious actor managed to redirect all communication to a malicious cloud.
The second method is simply intercepting the communications from a single robot to individually target a household. The traffic could be selectively overwritten to re target the robot to replace very specific ingredients. This would be very hard to detect until something happened.
The third is to penetrate the robot of a individual household and task it directly. Which if executed would be very similar to having a butler or maid in your house with malicious intent.
Encryption will be critical to securing these systems. Also building a robust cyber capability within the united states to protect citizens from devastating attacks such as this. Corporations will need to act in their self interest to ensure that the security of their systems is beyond questions. And then quietly internally consistently run penetration testing to ensure their systems are secure.
It no longer seems like science fiction to imagine a kitchen robot that would prepare a meal or food item, the governing AI will rely upon a hierarchy of understanding most likely housed in the cloud. This is a necessity because when a robot receives instructions to make a cake the variety of what flour can look like through every kitchen in the country requires an understanding of flower that transcends simply reading the table. This transcendent understanding of what flower is will likely be queried from a cloud of the manufacturer where the robot will receive the concept of what it is looking for.
The potential for malicious actors to manipulate this exists in three distinct methods. First the broad approach, where the entire cloud is penetrated and every system is affected. This will be quickly identified and quickly remedied. This is also very similar to what would happen if a malicious actor managed to redirect all communication to a malicious cloud.
The second method is simply intercepting the communications from a single robot to individually target a household. The traffic could be selectively overwritten to re target the robot to replace very specific ingredients. This would be very hard to detect until something happened.The third is to penetrate the robot of a individual household and task it directly. Which if executed would be very similar to having a butler or maid in your house with malicious intent.
Encryption will be critical to securing these systems. Also building a robust cyber capability within the united states to protect citizens from devastating attacks such as this. Corporations will need to act in their self interest to ensure that the security of their systems is beyond questions. And then quietly internally consistently run penetration testing to ensure their systems are secure.
