“The issues with Autonomy related to how the software recognised punctuation, the people said” - if folk knew how software operates they would be so much more prudent with investing in and relying on it 2/2

Serious Fraud Office reviews cases after software problems via Financial Times 1/ on.ft.com/3Uo7QpI

This poses some interesting questions both for the speech act theory of law and for ideas around legal materiality (how law is made 'real' in the world)

What if universities responded to AI hype with confidence in their core mission rather than FOMO?

Language models naturally generate outputs towards the 'center' of the data distribution they are trained on, neglecting rare/tail perspectives.

This paper argues that while useful, overreliance on recursive AI systems could lead to a 'knowledge collapse' phenomenon where public

Iris van Rooij 💭 mireillemoret @[email protected] at Mastodon Experts in LLM engineering and experts in legal reasoning typically have a different understanding of what a legal 'fact' is. This leads to a confusion on what can be automated. (see e.g. the CoHuBiCoL Typology of Legal Technologies & link.springer.com/article/10.100… by Masha Medvedeva)

matt Michael Veale @[email protected] Thank you for the open version. With apologies for the self-promotion, the 'self-driving laws'/'rules as code' aspects of this are ones that I have critiqued in a recent article in Law, Innovation & Technology: tandfonline.com/doi/full/10.10…

Truly incredible to produce a notably shallow and naive column after decades of writing and engagement with this topic.

The breathless enthusiasm of some senior members of the judiciary in England and Wales for the use of AI in the judicial sector is nothing short of extraordinary.

Integrating generative AI in this vision of education shouldn’t be that hard 🤓

Each of the suggestions in this article is fraught with potential difficulty. It would be prudent to sketch the challenges and potential issues to at least be aware of -- there is no shortage of literature on the subject (cc @cohubicol1)

My new article 'ChatGPT, Artificial Intelligence (AI) Large Language Models, and Law',

upcoming in the Fordham Law review,

is now posted on SSRN for those who are interested:

papers.ssrn.com/sol3/papers.cf…

Short paper from European Commission for the Efficiency of Justice (CEPEJ) on the use of generative AI by the judiciary: a basic outline of the potential problems and some simple guidance on when (not) to use it

coe.int/en/web/cepej/-…

Today I had the pleasure of delivering a presentation on the EU #AI Act and its applicability to #cybersecurity and critical infrastructure to representatives from respective Ministries and cybersecurity entities of Western Balkans organized by the DCAF HQ in Durres, Albania

I do not believe human-level AI (artificial superintelligence, or the commonest sense of #AGI ) is close at hand. AI has made breakthroughs, but the claim of AGI by 2030 is as laughable as claims of AGI by 1980 are in retrospect. Look how similar the rhetoric was in LIFE in 1970!

Jonathan Jones Sir Geoffrey Vos has been playing with his computer toys again… he’s an expert on AI now apparently 🤷😔👇

x.com/JoshuaRozenber…

Jonathan Jones Oddly the LLM’s Geoff consulted omitted to mention the work of CoHuBiCoL in this area… perhaps he could try consulting Pauline McBride & Dr Laurence Diver about their work 🤞🤷

Let's hope the political rush to bury Rule of Law under Rules as Code is subverted somewhat by the more nuanced approach advocated here. And yes, multi-interpretability and contestation are features not bugs