Remember in 2016, when the FBI took Apple to court over its refusal to provide assistance unlocking an iPhone? The device belonged to Syed Farook, the San Bernardino shooter whose rampage left 14 people dead. Apple’s contention was that complying with the request would require writing backdoor software that would serve as a “master key” – a dangerous precedent that could allow the FBI to access any iPhone. Ultimately, the case was dropped when the government said it had found a third party able to provide the service that Apple wouldn’t.
This was a high-profile case over the issue of data privacy, but there are many other under-the-radar investigations that are authorized by the courts or requested by law enforcement; in order to protect the integrity of each case, these investigations are kept secret while they are ongoing. Oftentimes charges never materialize and the sealed investigations are forgotten, but public accountability is severely compromised as a result.
That’s why researchers from CSAIL, the Computer Science and Artificial Intelligence Laboratory, and IPRI, the Internet Policy Research Initiative, both at MIT, have proposed a new cryptographic system designed to improve accountability while maintaining enough confidentiality to support the viability of investigations. The system is called AUDIT: Accountability of Unreleased Data for Improved Transparency.
AUDIT is designed to serve as a public ledger for data requests. Whenever a judge issues a secret court order or a law enforcement agency secretly requests data from a company, they are required to make a “cryptographic commitment” to later making the data request public.
The system can also be used to gauge whether actions by law enforcement agencies are consistent with what a court order actually allows, using a decades-old cryptographic method known as zero-method proofs that can prove proper surveillance conduct without revealing specific surveillance information.
In addition, statistical information can be aggregated so that the extent of surveillance can be studied at a larger scale. The research team developed a scalable aggregation system using an approach known as multi-party computation (MPC), which allows courts to disclose information without revealing their internal workings.
CSAIL graduate student Jonathan Frankle, one of the lead authors of a new paper about the system, noted that AUDIT could also be applied to any process in which data must be both kept secret and subject to public scrutiny. Clinical trials of new drugs, for example, often involve private information but also require enough transparency to assure regulators and the public that proper testing protocols are being observed.
"[AUDIT] represents a plausible way, both legally and technologically, for increasing public accountability through modern cryptographic proofs of integrity," said Eli Ben-Sasson, a professor in the computer science department at the Technion Israel Institute of Technology.
The team next plans to tweak the design of AUDIT in order to allow even more complex data requests to be handled. The possibility of partnering with working federal judges to develop a prototype for real-world use is also being explored.
"Lessons learned here will undoubtedly smooth the way towards greater accountability for a broader class of secret information processes, which are a hallmark of our digital age," said Stephen William Smith, a federal magistrate judge who has written extensively about government accountability.