1207, 2024

Reminder: Call for Papers DeepSec/DeepINTEL is still open!

René Pfeiffer/ July 12, 2024/ Conference/ 0 comments

It’s this time of the year again where the hot weather and deadlines collide. The call for papers for both DeepSec and DeepINTEL is still open! We are looking for original content, your creative ideas, and your invaluable experience. Please submit your proposal to our CfP manager. As always, we have a variety of topics we are interested in. The wonderful world of „artificial intelligence“ has taken the world and its CO2 output by storm. Large Language Models (LLMs) have „learned“ the Internet multiple times. Companies offering their LLM-based services promise to solve all kinds of tasks. What does this mean for IT security? Disinformation and propaganda are a big topic. Europe has already seen elections where structural disinformation played (and plays!) a vital role. Using false information in order to influence the voting

Read More

1207, 2024

IT Security, Standards, and Compliance

René Pfeiffer/ July 12, 2024/ Call for Papers, Conference, Legal/ 0 comments

You can often see the classic divide between technical and compliance persons in information technology within teams or organisations. Writing guidelines and writing configurations for implementation seem very different, with no overlaps. In reality, everyone has procedures. While they might not be written or follow a standardized format, having your ways of doing things is crucial to succeed in IT. The same goes for security. Creating policy documents and describing procedures in a way that technical minds can actually use them is a challenge. There is a crossover with the profession of writers who are experts in conveying nonfiction stories. And this is the origin of the schism between technicians and the compliance world. Badly written policies are a security risk, because no one takes them seriously. The purpose of your procedure documentation is

Read More

1906, 2024

Creative Writing is beneficial for Information Security

René Pfeiffer/ June 19, 2024/ Security, Stories

Do you like a good movie? Do you enjoy a good book? Your favourites most probably began as a piece of writing. There is a surprising overlap between creative writing, writing code, doing mathematics, and enjoying a well-defined information security configuration. Everything meets at the written word. IT documentation has a terrible reputation. Since it is always one or more steps behind the actual configuration, people prefer reading configurations instead. The magic is to keep changes in sync with your documentation. Another ingredient is to write concisely and to create the right structure. While documenting IT infrastructure is not like writing a script for a movie, it requires describing everything in the right order. You need to make sure people can look things up and find systems and controls required for security. An IT

Read More

1706, 2024

Online Security is threatened by Politics in the EU

René Pfeiffer/ June 17, 2024/ Communication, Security

A vote for the EU Chat Control disaster is scheduled again. If the vote passes, then this chat control system will destroy the security of communication platforms for all Internet users, organisation, and companies. Chat control basically disables end-to-end-encryption (E2EE). The proposals have been widely discussed, but it seems that the Belgian presidency decided to rush the topic right before the Summer break. Secure communications platforms such as Signal and Threema warned lawmakers. EuroISPA has published a statement of protest as well. E2EE is a cornerstone of information security. The chat control regulation would only affect citizens and businesses. Criminals will always have a way around this. The proposed mass surveillance is a security risk by itself. Nation state conducting espionage will welcome the weakening of the security posture and use it against the

Read More

0906, 2024

DeepSec 2024 Press Release: State Attacks on Information Security continue unabated. End-to-end Encryption remains an important and threatened Component of Security.

Sanna/ June 9, 2024/ Press

The introduction of strong encryption has repeatedly led to disputes with authorities and the government in the past. Whether it’s mobile networks, email systems, messengers or the World Wide Web, every iteration of the technical protocols requires backdoors that jeopardise the entire communications infrastructure. The DeepSec conference warns against opening the door to espionage. Secure or insecure, that is the Question Encryption inevitably has to do with mathematics, and the algorithms used in encryption technologies almost always originate from mathematical research. There are ready-made and well-tested components for IT infrastructures that are freely available. The critical point in securing communication is always to prevent messages from being intercepted. The only way to do this is with end-to-end encryption (EE2E). The keys involved remain exclusively with the sender and recipient. All parties involved in forwarding

Read More

0406, 2024

DeepSec 2024 Press Release: The limits of ‘AI’ language models lie in security. DeepSec warns: ‘AI’ language models generate content and override authorisations

Sanna/ June 4, 2024/ Conference, Press

    Language model algorithms, also known as generative artificial intelligence, continue to celebrate their supposed triumphant advance through many media platforms. Security researchers have analysed the products and revealed a number of weaknesses in the ‘AI’ applications. This year’s DeepSec conference is dedicated to the threats posed by ‘AI’ learning models that use incomplete restrictions to analyse public and sensitive data. Large Language Models (LLMs) as Auto-Completion The technical description of the many ‘artificial intelligence’ (‘AI’) products on the market is impressive. In simple terms, the concept behind the advertising campaigns consists of algorithms that copy as much data as possible, break it down and then recombine it to provide answers to any questions. The learning process when creating the language model is not initially monitored or moderated. Only in later phases does

Read More

2505, 2024

Science Fictions meets Large Language Models

René Pfeiffer/ May 25, 2024/ Conference

Given the advertising of the manufacturers using a Large Language Model (LLM) is just like having a conversation with a person. The reality looks different. Google’s AI search has recently recommended to glue pizza together, eats rocks, or jump off the Golden Gate bridge when being depressed. This is clearly bad advice. Apparently these answers are or were part of the learning process. Incidents like this and the hallucinations of LLM algorithms have been discussed already. Science fictions fans will recall conversations with computer or AI simulations where someone tries to trick the machine do override security checks. The Open Worldwide Application Security Project (OWASP) created a list of threats to LLM applications. The target audience are developers, designers, architects, managers, and organizations. Marc Pesce wrote an article about tests with different LLM implementations.

Read More

3003, 2024

Ross Anderson has died

René Pfeiffer/ March 30, 2024/ High Entropy, Misc

We mourn the loss of security researcher Ross Anderson. His contribution to information security and digital privacy was significant. His book Security Engineering is a cornerstone for everyone diving into the complex world of digital security. Ross taught at the University of Cambridge and Edinburgh University. He was a very prolific writer and published insights into the technological aspects of information security. He will be missed.

2903, 2024

DeepSec and DeepINTEL 2024 Call for Papers is open

René Pfeiffer/ March 29, 2024/ Call for Papers

The call for papers is open! DeepSec and DeepINTEL are waiting for your input. We are looking for your talks and trainings. Tell us what you found and tell our trainees how to defend against attacks. Please submit proposals for trainings as early as possible. We try to fill at least half of the trainings slots before the Summer, so interested persons have some more time to plan their attendance. Our main aim for 2024 is to examine the weaknesses of Large Language Models (LLMs) and explore their potential for exploitation. The obvious way is to use the prompt, but there are ways to influence of poison the training data. We have seen publications and nascent source code doing this. The less obvious way of weaponising these algorithms is to spread disinformation. Generated content

Read More

0403, 2024

Memory Safety revisited

René Pfeiffer/ March 4, 2024/ Conference

Memory safety is the most important problem in information security. This is something the White House and the NSA want you to believe. The recommendation is to use a different programming language, and all our problems will magically disappear. The proposal sounds a lot like the typical magical bullet solution, just like one of the many marketing promises of vendors since the 1990s. Attacks on memory buffers is the least of your current problems. Attackers use „living off the land“ attacks which use memory-safe scripting languages. If you look at the CWE statistics, then there are lots and lots of input validation errors that will bring down the security of many applications. Most web applications use questionable frameworks that are neither mature nor well-tested. Access to storage systems (SQL or NoSQL) still feature injections

Read More

0602, 2024

Encryption refreshed, Plans for 2024

René Pfeiffer/ February 6, 2024/ Conference

Computer science is all about automation. Repetitive tasks are best done by machines. This is true for our TLS certificate, but maybe you noticed it expired a few days ago. As always, this was because of an automated task that didn’t do what it was supposed to do. We changed parts of our infrastructure, so a few lines of code were not running on the new hardware. Blame it on ChatGPT, but your browser can trust our certificate again. Last year’s DeepSec conference had a focus on the zoo of artificial intelligence algorithms. The AI revolution has so far only pushed the Large Language Model (LLM) algorithms and a discussion about copyright. The battlefield is real. Researchers from the University of Chicago have published the Glaze and Nightshade algorithms to counter unrestricted harvesting by

Read More

3112, 2023

DeepSec wishes a Happy New Year 2024!

René Pfeiffer/ December 31, 2023/ Administrivia, Communication

We have been radio-silent throughout the December, because post-processing DeepSec and DeepINTEL 2023 took longer than usual. For everyone doing system maintenance, the month of December is also a wonderful opportunity to do some work behind the scenes. Thanks to the security bulletin about SMTP smuggling, there were some additional workarounds that needed attention. There will be another pause until we announce the next call for papers in February 2024. We have not yet decided on the focus. If you have some ideas, let us know. Enjoy the quiet days, and have a good transition into the new year 2024!

2411, 2023

Thanks for attending DeepSec and DeepINTEL 2023!

René Pfeiffer/ November 24, 2023/ Administrivia, Conference, DeepIntel

DeepSec 2023 ended a week ago, and it was amazing! We shout out a big thanks to all the speakers and all the attendees that made the conference memorable! Usually there is a period of several days after the conference where you hear nothing from us. We are not hibernating; we are in full post-production mode. Office life has caught up. The video material is currently being prepared for upload. Everyone who attended the conference will get early access to the presentations. Bear with us. We will send a notification once everything is ready. For everyone who missed the closing presentation, here are the dates for our events in 2024. Open your calendar, mark the dates. Also, do not forget to book early! We have a limit because of the conference venue safety regulations.

Read More

1511, 2023

DeepSec 2023 Talk and Breakout Session: Let’s Prepare for the Unexpected – Erlend Andreas Gjære

Sanna/ November 15, 2023/ Conference

What happens when a large group of more or less connected individuals need to deal with a cyber incident, together? In this interactive hands-on session, we will try to experience – first-hand – just how challenging it can be to keep information flowing, make the right decisions and protect our assets while dealing with a simulated crisis. We asked Erlend a few more questions about his talk and breakout session. Please tell us the top 5 facts about your talk and workshop. This will be an interactive session, and everyone can join! We are going to prepare for a cyber incident, together People share anonymous inputs via their phones Participants also receive individual updates on their phones There will be a breakout session afterwards for a deep-dive tabletop How did you come up with

Read More

0811, 2023

DeepSec 2023 – ENOMEM/EFBIG – Tickets sold out!

René Pfeiffer/ November 8, 2023/ Administrivia, Conference

This is the first time for us. The tickets for attending DeepSec on-site at the conference hotel are exhausted. We have no room to spare. You can only book training tickets (i.e. training without the conference) or tickets for accessing the live streams. Existing orders are still valid and will be processed. We have to take this step, because the space at the conference hotel gets too crowded. Furthermore, we have some limits regarding event security, and contrary to cloud platforms, we cannot sell more capacity than we have. Please consider accessing the live streams if you want to follow the presentations. You will also have the means to comment and ask questions. The stream access will also give you full access to all the recordings once we finished post-processing.