Law for the autonomous future
What happens to law as increasingly autonomous, self-executing technologies become embedded in our everyday lives? Our worry at Lautonomy is that accessing human rights and holding those with responsibility to account - whether governments, companies or individuals - becomes ever more complex ... and expensive.
To prevent this we're developing the tools needed to make the legal services industry fit for the autonomous future. Our goal is to ensure everybody, everywhere has the means to access the legal protections and remedies available.
We're in stealth mode for the moment but if you're interested in our aim and think you could help do drop us a line.
A foundational principle of law is that anyone should be free to access law's protections. The reality is that as legal systems have evolved into a complicated morass of jurisprudence access is mediated by lawyers and, hence, money. The situation is even more complicated in the global South where legal systems are generally weaker or less well established. Frontier technologies further deepen the challenge by making it exponentially harder - and more costly - to identify, track and attribute responsibility for wrongdoing.
In short, law has a complexity problem.
But we're in the AI age now, and perhaps the most radical thing about this is our emerging capability to solve for complexity and all its costs. That's essentially what we're setting out to do, to remove the complexity that systemically blocks fair and equitable access to law.
Our solution has two parts. What we can do now, today, and what we're building towards for tomorrow as converging technology improves.
In the now, we're creating a legal knowledge graph. This is essentially a map of the underlying structure of legal systems. With this we'll be able to do things like predict legal outcomes and legal risk, identify suspected violations and violators, and programmatically advise on the best legal process to follow to get redress.
In the then, we'll be able to use this knowledge base to generate fully formed legal arguments, spanning multiple issue areas and tailored for maximum effect in any jurisdiction. Perhaps even more fundamentally, we want to leverage an understanding of legal risk to change how human rights litigation gets financed.
What impact can we have? Consider financial crime and corruption, violations of which often have deep human rights implications. At present only 1-3% of global financial crimes are enforced against. That's a failure rate of 97-99%.
We believe in law's potential to create a better, more resilient, more human-kind world. We also believe that in too many countries and across too many issue areas this potential is at best locked up and at worst actively warped to benefit those with power. Law's failure rate actively undermines hopes for a more just, more equitable, more human-kind future.
What happens if Lautonomy can change that failure rate, even a little? It becomes easier to prosecute graft and corruption, making communities more secure. It becomes easier to secure property rights, spurring innovation. It becomes easier for companies and countries to enact progressive rules designed to end modern slavery, sanction perpetrators of mass atrocity and ensure ethical, responsible trade.
We're being ambitious, there's no getting away from that. But we also happen to think that given the pace of technological change ambition is a smart strategy.
We've concentrated our initial work on two particular forms of systemic harm: human rights and environmental harms.
In partnership with Freedom From Torture and the AIRE Centre we're looking at ways to improve redress and access to legal representation for torture survivors and asylum seekers. This is in a context where governments are beginning to use algorithmic methods to process asylum claims. You can read more here.
We're also beginning to look at environmental harms and ways that Lautonomy can support and strengthen emerging environmental accountability regimes. We'll be developing this track alongside the IEEE EarthLab.
Who's building Lautonomy? The core team is led by Dr. Andrew Jillions. We're also lucky enough to be able to draw on the expertise of Veronika Fikfak (University of Copenhagen; Human Rights Nudge Project); Annette So (Death Penalty Project); Ben Hayes (AWO); Olivia Belton (Leverhulme Centre for the Future of Intelligence, Cambridge University); Didem Gurdur Broo (Cambridge University); Tracy Doig (Freedom from Torture); Christopher Burr (University of Oxford; Alan Turing Institute); Nicolas Economou (Athens Roundtable on AI and the Rule of Law; H5); Kyriakos Kantarakias (Imperial College London); Kolbassia Haoussou (Survivors Speak Out; Freedom from Torture); Anthony F. Lang, Jr. (University of St. Andrews); Paco Nathan (Derwen.ai); Markella Papadouli (AIRE Centre); Michael Pereira (Mobilization for Justice); Swapnil Singh (National Law University, Lucknow); Azure Wheeler (Open Society Foundation); Akin Agunbiade (Digilaw); Colum Elliot-Kelly (Filament AI).
We're actively growing the project team so if you have relevant skills do get in touch. In particular we'd love to hear from anyone with a background or interest in knowledge graphs and natural language processing.
We'd love to hear from you. If you're interested in what Lautonomy is working towards do get in touch.
We’ve set ourselves the challenge of improving the global anti-torture regime.
The prohibitions against torture prohibits states or their agents from using torture; establishes a state obligation to investigate and prosecute instances of torture and prohibits the return (refoulement) of individuals to another state where they would be at risk of torture. The linked rules around cruel, inhuman or degrading treatment (or CIDT, for short) establish a wider range of obligations on states especially around the conditions in which people are held or forced to live.
Despite the wide support for these anti-torture rules, the practical, systemic protection and promotion is regularly fails.
In part this is a reflection of the nature of torture and CIDT. These are acts which are inherently difficult to prove, taking place behind closed doors and in secret, used or otherwise supported by powerful actors to silence criticism, sustained with threat of further harm to the individual and their families, and justice reliant on victims who often suffer psychological trauma associated with their experiences can have their evidence dismissed as “lacking credibility”. Refugees and asylum seekers attempting to claim rights of non-refoulement can be frustrated by a reliance on what are called "diplomatic assurances", essentially a non-justiciable promise by a receiving state not to torture any returnees, or face other difficult-to-meet and inconsistently applied evidentiary standards. Practical barriers to getting legal redress - including the financial and political risks individuals bear in demanding redress - also conspire to hollow out the transformative potential of the rules prohibiting torture and CIDT.
What can we do about this?
Lautonomy is working with Freedom From Torture and The AIRE Centre to help individuals understand, at a glance, the rights they as an individual are able to claim, the legal pathways to redress available, the likely evidence that will be needed to support their claim, and any useful supplementary information or analysis - such as academic research. Where there is an actionable case we aim to connect the individual with the best possible legal representation and, if necessary, access any funding required.
Mapping torture law
Good AI systems start with a good dataset. To get to actionable insights regarding torture law we need to be able to model a deep understanding of three core things or entities representing:
1. Knowledge of an individual’s identity. What is their citizenship, sexuality, gender, religion, ethnicity, or political affiliation? What work have they or do they do? Do they have family relationships or social networks that may have factored into their ill treatment? What is it that led to their being tortured or at risk of torture?
2. Knowledge of an event. What happened? What forms did the abuse take? Where did the events occur? Who were the perpetrators? What legal processes are they currently dealing with or fighting against?
3. Knowledge of the law. What rights can a particular individual access? What legal principles or jurisprudence governs in the particular circumstances of an individual’s experiences and identity? What legal strategy or process would be most appropriate given the facts of a particular case? What evidence will judges or state officials ask for?
Once we have these entities mapped out we can begin to ask questions in much the same way as a judge would, looking for the fit or correlation between an individual’s identity, a narrative of events and an understanding of the governing law and jurisprudence.
More to follow!
Interested in being involved with strengthening the anti-torture regime? Get in touch.