AI Transparency not doing so well

Posted on by

Stanford University researchers just released a report in which a “Foundation Model Transparency Index” (here) is presented. The first evaluation did not go so well, since the highest score is 54 out of 100. Comments by reviewers and experts in the field point out that “transparency is on the decline while capability is going through the roof” as Stanford CRFM Director Percy Liang told Reuters in an interview (see also here).

Post Quantum Cryptography on the Rise

Posted on by

Recently there have been quite a few announcements about the adoption of Post Quantum Cryptography algorithms to supplement, not yet substitute, current algorithms so that encryption will be able to withstand also when (and if) Quantum Computers will arrive. Latest is the announcement by Signal, as it has been reported for example here.

Compliance of Foundation AI Model Providers with Draft EU AI Act

Posted on by

Interesting study by the Center for Reasearch on Foundation Models, Stanford University, Human-Centered Artificial Intelligence, on compliance of Foundation Model Providers, such as OpenAI, Google and Meta, with the Draft EU AI Act. Here is the link to the study, and the results indicate that the 10 providers analysed “largely do not” comply with the draft requirements of the EU AI Act.

AI and the Extinction of the Human Race

Posted on by

The Center for AI Safety (here) has just published the following “Statement on AI Risk” (here):

Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.

The list of signatures is impressive (just look at the first 4) and should make us think more deeply about us and AI & ML.

On Large Language Models (and AI Models) Explainability

Posted on by

Researchers at OpenAI have recently released a scientific paper (here) entitled “Language models can explain neurons in language models“. The paper is quite technical, but it is interesting to quote from the Introduction:

Language models have become more capable and more widely deployed, but we do not understand how they work. Recent work has made progress on understanding a small number of circuits and narrow behaviors, but to fully understand a language model, we’ll need to analyze millions of neurons. This paper applies automation to the problem of scaling an interpretability technique to all the neurons in a large language model. Our hope is that building on this approach of automating interpretability will enable us to comprehensively audit the safety of models before deployment.

and to read the concluding Discussion section.

Windows code and Rust

Posted on by

Quite an interesting news (here for example): Microsoft is rewriting core Windows libraries in the Rust programming language instead of C/C++. Rust is a programming language which is “memory safe” which means that prevents entire classes of bugs which can lead to vulnerabilities and exploitations.

Do not expect that all Windows will be rewritten in Rust, it would be an enormous task and probably it will make little practical and technical sense. But rewriting key components of the operating system in a security minded programming language is for sure a great security step forward.