#mlsec
At BIML, we have been talking about recursive pollution and model collapse since 2020. Need some evidence? Look no further than search...
https://www.theregister.com/2025/05/27/opinion_column_ai_model_collapse
Why so much prompt injection in AI? 1. We don't follow the security engineering design principle "economy of mechanism," and 2, input to LLMs mixes control and data with impunity. We know better. #MLsec #infosec #security
https://www.darkreading.com/vulnerabilities-threats/llms-on-rails-design-engineering-challenges
@patrickcmiller absolute bullshit from anthropic whose anthropomorphic hogwash is a standard trick
See, for example, https://berryvilleiml.com/2024/02/08/absolute-nonsense-from-anthropic-sleeper-agents/
THIS IS BULLSHIT MULTIPLIED BY BAD REPORTING #MLsec
For the record and for history ...you cannot secure #AI by red teaming with known attacks. Just as with #swsec, you must build security in, starting with design.
Penetrate and patch is a braindead paradigm. Let's do some real #MLsec please.
https://www.scworld.com/news/nist-releases-new-ai-attack-taxonomy-with-expanded-genai-section
"Just let AI do it," they said. They being Microsoft. ROFL.
#MLsec #swdev #developer
Why AI should not do journalism...even "advertorials" which are already bullshit. #MLsec
https://arstechnica.com/ai/2025/05/chicago-sun-times-prints-summer-reading-list-full-of-fake-books/
This notion that early-career coders can be replaced by AI is wrong. Nobody is thinking about maintaining software or the architectural disaster to come when inscrutable code that mostly works is all over the codebase. And it's real..."early-career coders have been hit especially hard because much of what they do can now be done by AI."
I recommend hiring junior humans over AI. Still. Just tool them up.
https://www.wsj.com/lifestyle/careers/tech-jobs-hiring-artifical-intelligence-35cd66b0
Sometimes AI makes up shit about people...sometimes people make up shit about AI. In either case, beware the hype. #ML #AI #MLsec
https://www.wsj.com/tech/ai/mit-says-it-no-longer-stands-behind-students-ai-research-paper-11434092
BS does not scale in science land #ML #MLsec
https://www.wsj.com/tech/ai/mit-says-it-no-longer-stands-behind-students-ai-research-paper-11434092
Reviewing this absolute garbage work that has a veneer of science. What a joke. If this is the kind of #MLsec out of Europe that is supposed to save us, we're screwed.
Academic journals in security are utterly useless. #infosec #security
https://www.sciencedirect.com/science/article/pii/S0167404824002931
@baldur have you read our work? You might appreciate it. You can use it to shut those guys up.
One person's data pollution is another person's data gold. As long as we have next to zero insight into the immense training data sets used by LLMs this will happen again and again. Data protection fail groundhog day. #ML #AI #MLsec #security
https://www.darkreading.com/cyberattacks-data-breaches/deepseek-breach-opens-floodgates-dark-web
Repeat after me. You can't open test or red team AI into security. #MLsec
Benchmarks as popularity contests don't work. There are lots of other reasons that benchmarks have become almost worthless in #AI. In particular #ML benchmark as badnessometer comes to mind. #MLsec
https://techcrunch.com/2025/04/22/crowdsourced-ai-benchmarks-have-serious-flaws-some-experts-say/