Friday Squid Blogging: Baby Ichthyosaurus Fed on Squid

New discovery: paper and article. As usual, you can also use this squid post to talk about the security stories in the news that I haven’t covered. Read my blog posting guidelines here….

New discovery: paper and article.

As usual, you can also use this squid post to talk about the security stories in the news that I haven't covered.

Read my blog posting guidelines here.

from https://www.schneier.com/blog/

Hacking Voice Assistant Systems with Inaudible Voice Commands

Turns out that all the major voice assistants — Siri, Google Now, Samsung S Voice, Huawei HiVoice, Cortana and Alexa — listen at audio frequencies the human ear can’t hear. Hackers can hijack those systems with inaudible commands that their owners can’t hear. News articles….

Turns out that all the major voice assistants -- Siri, Google Now, Samsung S Voice, Huawei
HiVoice, Cortana and Alexa -- listen at audio frequencies the human ear can't hear. Hackers can hijack those systems with inaudible commands that their owners can't hear.

News articles.

from https://www.schneier.com/blog/

New Techniques in Fake Reviews

Research paper: "Automated Crowdturfing Attacks and Defenses in Online Review Systems." Abstract: Malicious crowdsourcing forums are gaining traction as sources of spreading misinformation online, but are limited by the costs of hiring and managing human workers. In this paper, we identify a new class of attacks that leverage deep learning language models (Recurrent Neural Networks or RNNs) to automate the…

Research paper: "Automated Crowdturfing Attacks and Defenses in Online Review Systems."

Abstract: Malicious crowdsourcing forums are gaining traction as sources of spreading misinformation online, but are limited by the costs of hiring and managing human workers. In this paper, we identify a new class of attacks that leverage deep learning language models (Recurrent Neural Networks or RNNs) to automate the generation of fake online reviews for products and services. Not only are these attacks cheap and therefore more scalable, but they can control rate of content output to eliminate the signature burstiness that makes crowdsourced campaigns easy to detect.

Using Yelp reviews as an example platform, we show how a two phased review generation and customization attack can produce reviews that are indistinguishable by state-of-the-art statistical detectors. We conduct a survey-based user study to show these reviews not only evade human detection, but also score high on "usefulness" metrics by users. Finally, we develop novel automated defenses against these attacks, by leveraging the lossy transformation introduced by the RNN training and generation cycle. We consider countermeasures against our mechanisms, show that they produce unattractive cost-benefit tradeoffs for attackers, and that they can be further curtailed by simple constraints imposed by online service providers.

from https://www.schneier.com/blog/

A Framework for Cyber Security Insurance

New paper: "Policy measures and cyber insurance: a framework," by Daniel Woods and Andrew Simpson, Journal of Cyber Policy, 2017. Abstract: The role of the insurance industry in driving improvements in cyber security has been identified as mutually beneficial for both insurers and policy-makers. To date, there has been no consideration of the roles governments and the insurance industry should…

New paper: "Policy measures and cyber insurance: a framework," by Daniel Woods and Andrew Simpson, Journal of Cyber Policy, 2017.

Abstract: The role of the insurance industry in driving improvements in cyber security has been identified as mutually beneficial for both insurers and policy-makers. To date, there has been no consideration of the roles governments and the insurance industry should pursue in support of this public­-private partnership. This paper rectifies this omission and presents a framework to help underpin such a partnership, giving particular consideration to possible government interventions that might affect the cyber insurance market. We have undertaken a qualitative analysis of reports published by policy-making institutions and organisations working in the cyber insurance domain; we have also conducted interviews with cyber insurance professionals. Together, these constitute a stakeholder analysis upon which we build our framework. In addition, we present a research roadmap to demonstrate how the ideas described might be taken forward.

from https://www.schneier.com/blog/

Proof that HMAC-DRBG has No Back Doors

New research: "Verified Correctness and Security of mbedTLS HMAC-DRBG," by Katherine Q. Ye, Matthew Green, Naphat Sanguansin, Lennart Beringer, Adam Petcher, and Andrew W. Appel. Abstract: We have formalized the functional specification of HMAC-DRBG (NIST 800-90A), and we have proved its cryptographic security — that its output is pseudorandom — using a hybrid game-based proof. We have also proved that…

New research: "Verified Correctness and Security of mbedTLS HMAC-DRBG," by Katherine Q. Ye, Matthew Green, Naphat Sanguansin, Lennart Beringer, Adam Petcher, and Andrew W. Appel.

Abstract: We have formalized the functional specification of HMAC-DRBG (NIST 800-90A), and we have proved its cryptographic security -- that its output is pseudorandom -- using a hybrid game-based proof. We have also proved that the mbedTLS implementation (C program) correctly implements this functional specification. That proof composes with an existing C compiler correctness proof to guarantee, end-to-end, that the machine language program gives strong pseudorandomness. All proofs (hybrid games, C program verification, compiler, and their composition) are machine-checked in the Coq proof assistant. Our proofs are modular: the hybrid game proof holds on any implementation of HMAC-DRBG that satisfies our functional specification. Therefore, our functional specification can serve as a high-assurance reference.

from https://www.schneier.com/blog/

Hacking a Gene Sequencer by Encoding Malware in a DNA Strand

One of the common ways to hack a computer is to mess with its input data. That is, if you can feed the computer data that it interprets — or misinterprets — in a particular way, you can trick the computer into doing things that it wasn’t intended to do. This is basically what a buffer overflow attack is: the…

One of the common ways to hack a computer is to mess with its input data. That is, if you can feed the computer data that it interprets -- or misinterprets -- in a particular way, you can trick the computer into doing things that it wasn't intended to do. This is basically what a buffer overflow attack is: the data input overflows a buffer and ends up being executed by the computer process.

Well, some researchers did this with a computer that processes DNA, and they encoded their malware in the DNA strands themselves:

To make the malware, the team translated a simple computer command into a short stretch of 176 DNA letters, denoted as A, G, C, and T. After ordering copies of the DNA from a vendor for $89, they fed the strands to a sequencing machine, which read off the gene letters, storing them as binary digits, 0s and 1s.

Erlich says the attack took advantage of a spill-over effect, when data that exceeds a storage buffer can be interpreted as a computer command. In this case, the command contacted a server controlled by Kohno's team, from which they took control of a computer in their lab they were using to analyze the DNA file.

News articles. Research paper.

from https://www.schneier.com/blog/

Friday Squid Blogging: Squid Eyeballs

Details on how a squid’s eye corrects for underwater distortion: Spherical lenses, like the squids’, usually can’t focus the incoming light to one point as it passes through the curved surface, which causes an unclear image. The only way to correct this is by bending each ray of light differently as it falls on each location of the lens’s surface….

Details on how a squid's eye corrects for underwater distortion:

Spherical lenses, like the squids', usually can't focus the incoming light to one point as it passes through the curved surface, which causes an unclear image. The only way to correct this is by bending each ray of light differently as it falls on each location of the lens's surface. S-crystallin, the main protein in squid lenses, evolved the ability to do this by behaving as patchy colloids­ -- small molecules that have spots of molecular glue that they use to stick together in clusters.

Research paper.

As usual, you can also use this squid post to talk about the security stories in the news that I haven't covered.

Read my blog posting guidelines here.

from https://www.schneier.com/blog/

Confusing Self-Driving Cars by Altering Road Signs

Researchers found that they could confuse the road sign detection algorithms of self-driving cars by adding stickers to the signs on the road. They could, for example, cause a car to think that a stop sign is a 45 mph speed limit sign. The changes are subtle, though — look at the photo from the article. Research paper: "Robust Physical-World…

Researchers found that they could confuse the road sign detection algorithms of self-driving cars by adding stickers to the signs on the road. They could, for example, cause a car to think that a stop sign is a 45 mph speed limit sign. The changes are subtle, though -- look at the photo from the article.

Research paper:

"Robust Physical-World Attacks on Machine Learning Models," by Ivan Evtimov, Kevin Eykholt, Earlence Fernandes, Tadayoshi Kohno, Bo Li, Atul Prakash, Amir Rahmati, and Dawn Song:

Abstract: Deep neural network-based classifiers are known to be vulnerable to adversarial examples that can fool them into misclassifying their input through the addition of small-magnitude perturbations. However, recent studies have demonstrated that such adversarial examples are not very effective in the physical world--they either completely fail to cause misclassification or only work in restricted cases where a relatively complex image is perturbed and printed on paper. In this paper we propose a new attack algorithm--Robust Physical Perturbations (RP2)-- that generates perturbations by taking images under different conditions into account. Our algorithm can create spatially-constrained perturbations that mimic vandalism or art to reduce the likelihood of detection by a casual observer. We show that adversarial examples generated by RP2 achieve high success rates under various conditions for real road sign recognition by using an evaluation methodology that captures physical world conditions. We physically realized and evaluated two attacks, one that causes a Stop sign to be misclassified as a Speed Limit sign in 100% of the testing conditions, and one that causes a Right Turn sign to be misclassified as either a Stop or Added Lane sign in 100% of the testing conditions.

from https://www.schneier.com/blog/

Detecting Stingrays

Researchers are developing technologies that can detect IMSI-catchers: those fake cell phone towers that can be used to surveil people in the area. This is good work, but it’s unclear to me whether these devices can detect all the newer IMSI-catchers that are being sold to governments worldwide. News article….

Researchers are developing technologies that can detect IMSI-catchers: those fake cell phone towers that can be used to surveil people in the area.

This is good work, but it's unclear to me whether these devices can detect all the newer IMSI-catchers that are being sold to governments worldwide.

News article.

from https://www.schneier.com/blog/