Like pretty much any other computer-controlled device, 3D printers can be maliciously hacked. Given that they're now creating items such as aircraft parts and medical implants, the results of a compromised print job could be catastrophic – particularly if the fault wasn't visible from the outside. That's why scientists from Rutgers University and the Georgia Institute of Technology have developed a new system for making sure that 3D printers are doing what they're supposed to.

The system operates on three levels.

First of all, it records the sound of the printer as it builds what is verified as a "correct" version of the piece in question. When other copies of the object are subsequently printed, recordings of those print jobs are compared to that initial reference recording. If there are significant differences, it indicates that the printer may be running malicious software.

Secondly, the system tracks the movements of the printer's components, such as its extruder. Again, if that movement pattern doesn't match what has initially been verified as the "right" pattern, then it follows that the printer may not be performing correctly.

Finally, tiny gold nanorods can be mixed in with the filament material. These don't affect the integrity of the finished product, and their locations within the item can be ascertained using Raman spectroscopy and computed tomography. In this way, hidden flaws such as internal holes can be detected via the positioning of the nanorods.

The scientists have already demonstrated that 3D printers can be hacked, plus they've successfully trialled their system on three different types of printers which were each building a polyethylene tibial knee prosthesis – something that definitely needs to be printed correctly.

The research will be presented this Friday at the USENIX Security Symposium in Vancouver.