A new radio identification (RFID) chip developed by American researchers could bring an end to identity theft.
Developed by researchers from the Massachusetts Institute of Technology (MIT) and electronics company Texas Instruments, the hack-proof chip prevents the two most common types of attacks designed to steal a cryptographic key – the so-called side-channel attack and the power glitch attack.
During a side-channel attack the hacker attempts to extract a cryptographic key by analysing patterns in memory access and fluctuations of power usage during a cryptographic operation. In a power glitch attack the goal is to circumvent limits on the number of incorrect password entries. The attacker cuts power supply to the chip right before it changes the secret key, making it impossible to complete the operation.
The MIT team led by Chiraag Juvekar addressed the first problem by designing the new chip to run a random-number generator that would provide a new secret key after every transaction. A central server would run the same generator and every time an RFID scanner queried the tag, it would relay the results to the server, to see if the current key was valid.
To make the power-glitch attacks impossible, the researchers designed the chip to contain its own power supply, since one of the major weaknesses of current RFID chips is that they are charged by tag readers and have no on-board power source.
The power supply was designed in a way that makes it almost impossible to cut. In addition to that, the chip is equipped with what the researchers call non-volatile memory cells. These cells store all the data the chip is using at the moment when it begins to lose power.
Several prototypes of the hack-proof chip have been built by Texas Instruments and subjected to rigorous testing. Presenting the research at the International Solid-State Circuits Conference in San Francisco, the researchers said the technology worked exactly as envisioned.
If such chips were widely adopted, it could mean that an identity thief couldn't steal a credit card number or key card information just by sitting next to the card user at a café and high-tech burglars couldn't swipe expensive goods from a warehouse and replace them with dummy tags.
At the San Francisco conference, another MIT team introduced a chip designed specifically to implement neural networks. Ten times as efficient as a mobile GPU, the chip could enable mobile devices to run powerful artificial-intelligence algorithms locally, rather than uploading data to the Internet for processing.
The chip, dubbed Eyeriss, could also speed up the development of the Internet of Things by enabling connected devices, including robots and autonomous cars, to make intelligent decisions by themselves only by running artificial intelligence algorithms.