Online dispute resolution for real: Ebay Spinoff Modria Is Judge Judy For Cyber Shoppers (Fast Company)
Case factor analysis for real: Judicata Raises $2M From Peter Thiel, Keith Rabois And Others To Give Lawyers Better Research And Analytics Tools (TechCrunch)
How To Run A Law Firm Like A Startup (Business Insider via Computational Legal Studies)
LegalForce Store Offers Walk-in Lawyer Access in Palo Alto (eLawyering Blog) (BTW, there has been something a bit similar called the Legal Lounge for a while now here in Helsinki as well.)
Pitchpitchpitch! ReInvent Law are arranging a pitch competition next week. An excellent initiative and the first one specifically for legal startups I've heard about. As a cynical old person who has hung around young and innovative startupsters a lot last year at the Startup Sauna I would just like to warn that while pitching is fine and useful for early-stage companies, it is easy to overdo and just keep pitchpitchpitching instead of working on the product. (Oh, and Onomatics got seed funding without even presenting the pitch deck. Neener neener.)
Results of said competition @ReneeKnake: @ReInventLaw StartUP Competition winners! Best overall LoquiTab Runner up @Kat_Hennessy RiskADvantage Most creative @A_Ninhja p(AR)adigmLaw
Musings on Law and Intelligence (Artificial and Natural)
"The future is already here - it's just not very evenly distributed." William Gibson
"Wisdom is the abstract of the past, but beauty is the promise of the future." Oliver Wendell Holmes
Wednesday, 6 February 2013
Wednesday, 16 January 2013
Translation Technology and Copyright
Translation technology (TT) has made significant technical advances over the past decade. Machine translation has become increasingly commonplace in everyday use through services such as Google Translate. At the same time, computer-aided translation systems are now an invaluable tool for the professional translator. Unlike the machine translation systems of yore, current systems are no longer based on explicit, formal models of language, but rather on machine learning and statistical methods using vast collections of multilingual documents with each language element aligned across language pairs. Because of this, entirely new questions of intellectual property become a crucial yet still poorly understood part of the enterprise. This article presents an introduction to translation technology and the legal questions potentially involved and then proceeds to the legal analysis of quantitative and qualitative requirements for copyright protection, the issue of potential (human and computer) authorship to translation technology output, and others.
Our article can be found here:
Sorry for the paywall. Our article has not undergone any language examination. According to the editors each text “shall be the voice of its author”.
Thursday, 13 December 2012
What is Artificial Intelligence?
Prompted by the discussion at a presentation I gave yesterday on intelligent legal technology (titlend Älykäs oikeusteknologia, ie. in Finnish, for once) I guess I feel the need to export a part of the neural network that subsequently emerged in the form of a blog post. I was asked to define artificial intelligence, and since I refused to provide a definition, I was asked again. And again. So even if providing such a definition is quite irrelevant as I don’t research legal AI in general (in which case the delineation between AI and non-AI might be of interest) but rather some specific questions (modelling vagueness and uncertainty in law) which without a question are AI & law questions, here, for explanatory rather than definitive use, with no warranties for fitness for any particular purpose, yadda yadda yadda, are my two cents:
Artificial Intelligence is the cross-disciplinary enterprise of trying to do things with a computer which when done by people are said to require intelligence and which computers cannot (yet) do. (The careful reader may notice a certain degree of isomorphism with a popular definition within the extended cognition framework...)
(And comparison shoppers, here is Wikipedia’s current version: “Artificial intelligence (AI) is the intelligence of machines and robots and the branch of computer science that aims to create it.”)
So: Consistently with the bottom-up approach to AI I like to advocate in general, I don’t think allusions to the Turing test or the Singularity or whatever are all that interesting, as far as actual progress is concerned, the cognitive arts advance through innovations which are very small increments from the perspective of AI as a whole but can be quite dramatic for the topical discipline in question.
I do think that the trying (or aim[ing] to create) is an important part of what makes AI AI. Doing arithmetics also requires intelligence but has never been a part of AI since computers could do (and indeed were built to do) it properly from the beginning. And so on the way from notrespondingstilltrying to commercial viability, AI projects start being called computational whatever or whatever technology (hence legal technology). Of course the boundaries are vague and the whole boxological excercise of little use in anything other than turf-wars in academia.
And the fact that the definition refers to human intelligence just serves to illustrate the futility and question-beggitude of definitions for one simple reason: The psychological understanding of human intelligence just adds even more layers of complexity. For example, IQ tests cannot possibly measure human intelligence per se and in general. What they measure instead is a specific indicator known as the g factor (or general intelligence), which has been shown to correlate (reasonably) well with the more specific intelligent abilities.
Even if working on definitions can occasionally serve an useful purpose, personally I think in most cases the more expedient alternative is to follow Justice Potter Stewart in Jacobellis v. Ohio: “I know it when I see it.” For historical reasons, jurisprudence in Finland still has a particular affinity for concepts and definitions not really seen elsewhere to the same degree. I’m planning to address this issue in extenso at some point with the title Begriffsjurisprudenz 2.0. (Spoiler alert: may also offend ontologists.)
Artificial Intelligence is the cross-disciplinary enterprise of trying to do things with a computer which when done by people are said to require intelligence and which computers cannot (yet) do. (The careful reader may notice a certain degree of isomorphism with a popular definition within the extended cognition framework...)
(And comparison shoppers, here is Wikipedia’s current version: “Artificial intelligence (AI) is the intelligence of machines and robots and the branch of computer science that aims to create it.”)
So: Consistently with the bottom-up approach to AI I like to advocate in general, I don’t think allusions to the Turing test or the Singularity or whatever are all that interesting, as far as actual progress is concerned, the cognitive arts advance through innovations which are very small increments from the perspective of AI as a whole but can be quite dramatic for the topical discipline in question.
I do think that the trying (or aim[ing] to create) is an important part of what makes AI AI. Doing arithmetics also requires intelligence but has never been a part of AI since computers could do (and indeed were built to do) it properly from the beginning. And so on the way from notrespondingstilltrying to commercial viability, AI projects start being called computational whatever or whatever technology (hence legal technology). Of course the boundaries are vague and the whole boxological excercise of little use in anything other than turf-wars in academia.
And the fact that the definition refers to human intelligence just serves to illustrate the futility and question-beggitude of definitions for one simple reason: The psychological understanding of human intelligence just adds even more layers of complexity. For example, IQ tests cannot possibly measure human intelligence per se and in general. What they measure instead is a specific indicator known as the g factor (or general intelligence), which has been shown to correlate (reasonably) well with the more specific intelligent abilities.
Even if working on definitions can occasionally serve an useful purpose, personally I think in most cases the more expedient alternative is to follow Justice Potter Stewart in Jacobellis v. Ohio: “I know it when I see it.” For historical reasons, jurisprudence in Finland still has a particular affinity for concepts and definitions not really seen elsewhere to the same degree. I’m planning to address this issue in extenso at some point with the title Begriffsjurisprudenz 2.0. (Spoiler alert: may also offend ontologists.)
Saturday, 8 December 2012
Peter Thiel on Singularity and legal technology
Betabeat (just one of many tech blogs I follow regularly) had yesterday a interesting post on Peter Thiel’s (Stanford Law graduate, PayPal co-founder &c &c) presentation at the Legal Technology and Informatics course held at Stanford law for the first time this past autumn. Blake Winters has kindly written and published an essay on the presentation, on which these brief comments are based.
Personally I think all this talk about the Singularity is mostly just a distraction (and of course fodder for dystopic science fiction). Actually functioning general-purpose artificial intelligence is not simply just a matter of bytes and CPU cycles or even fully replicating the neural network of a human brain at some instant (because so much of human intelligence depends on neurogenesis and the formation and pruning of connections, processes which only a couple of decades ago were still thought to end by adulthood), and anyway it is so far in the horizon that it is impossible to use as a target. There is still a lot of work to be done in trying to make sense about the actual functioning of human cognition. (The discussion about free will and whether Libet’s experiments show that it doesn’t exist is a good example.) Even if the Singularity does arrive at some point, the interaction of humans and computers at that time will not be something we can easily imagine. (Just compare whatever you are using to read this with a completely character-based interface (your only choice thirty years ago). And I still fondly remember the sound of a good mechanical teleprinter...)
To date, AI has been most successful when trying to solve very difficult but still quite concrete problems with computational methods. My rule of thumb is that when AI starts being useful, it stops being called AI. (Hence I also prefer to talk about (intelligent) legal technology rather than legal AI.) There are many branches of computer science and other computational sciences which started out basic AI research, with language technology as just one good example.
But more importantly, as for the shorter timeframe, I totally agree with Thiel. Computers are much better than people at some tasks and legal technology has great potential for radically transforming the marketplace for legal services (for the better) in the near future. The work we do at Onomatics will hopefully a good example from the more technologically advanced end of the scale, but our domain (trademark law) is just a very small corner of the entire legal system.
All this just reminds me that I should finally get around to writing two blog posts I’ve been thinking about for quite a while, one titled “Why do computers make better lawyers than people” and the other – of course – “Why do people make better lawyers than computers”. Real Soon Now!
Further reading:
Personally I think all this talk about the Singularity is mostly just a distraction (and of course fodder for dystopic science fiction). Actually functioning general-purpose artificial intelligence is not simply just a matter of bytes and CPU cycles or even fully replicating the neural network of a human brain at some instant (because so much of human intelligence depends on neurogenesis and the formation and pruning of connections, processes which only a couple of decades ago were still thought to end by adulthood), and anyway it is so far in the horizon that it is impossible to use as a target. There is still a lot of work to be done in trying to make sense about the actual functioning of human cognition. (The discussion about free will and whether Libet’s experiments show that it doesn’t exist is a good example.) Even if the Singularity does arrive at some point, the interaction of humans and computers at that time will not be something we can easily imagine. (Just compare whatever you are using to read this with a completely character-based interface (your only choice thirty years ago). And I still fondly remember the sound of a good mechanical teleprinter...)
To date, AI has been most successful when trying to solve very difficult but still quite concrete problems with computational methods. My rule of thumb is that when AI starts being useful, it stops being called AI. (Hence I also prefer to talk about (intelligent) legal technology rather than legal AI.) There are many branches of computer science and other computational sciences which started out basic AI research, with language technology as just one good example.
But more importantly, as for the shorter timeframe, I totally agree with Thiel. Computers are much better than people at some tasks and legal technology has great potential for radically transforming the marketplace for legal services (for the better) in the near future. The work we do at Onomatics will hopefully a good example from the more technologically advanced end of the scale, but our domain (trademark law) is just a very small corner of the entire legal system.
All this just reminds me that I should finally get around to writing two blog posts I’ve been thinking about for quite a while, one titled “Why do computers make better lawyers than people” and the other – of course – “Why do people make better lawyers than computers”. Real Soon Now!
Further reading:
- Peter Thiel on The Future of Legal Technology - Notes Essay (by Blake Masters)
- Notes from Peter Thiel’s CS183 Startup class at Stanford (by Blake Masters)
- Syllabus for the Stanford Legal Technology and Informatics course
- Course reader for the same
Thursday, 8 November 2012
Happy World Usability Day!
Today, November 8th, is World Usability Day, arranged annually since 2006. List of events here (although if you are interested in events near you, the map on the front page is much more usable), there’s even one here in Helsinki).
If you only know usability from real-world usability, or more likely the lack of it (good usability is unobtrusive), here are the standard definitions:
Of course I’ll also take the opportunity to mention my paper titled Software Usability and Legal Informatics (draft paper on SSRN) which I will be presenting later this month at the KnowRight conference. As far as I know, there has been very little earlier scholarship on the topic in legal informatics, but pointers are most welcome. I will be pursuing this line of research further in other articles at least over the next couple of years.
[Update: presentation now available here.]
If you only know usability from real-world usability, or more likely the lack of it (good usability is unobtrusive), here are the standard definitions:
“[Usability refers to] the extent to which a product can be used by specified users to achieve specified goals with effectiveness, efficiency and satisfaction in a specified context of use.” - ISO 9241-11For more information, see The User Experience Professionals’ Association website.
“Human-centered design is characterised by: the active involvement of users and a clear understanding of user and task requirements; an appropriate allocation of function between users and technology; the iteration of design solutions; multi-disciplinary design.” - ISO 13407
Of course I’ll also take the opportunity to mention my paper titled Software Usability and Legal Informatics (draft paper on SSRN) which I will be presenting later this month at the KnowRight conference. As far as I know, there has been very little earlier scholarship on the topic in legal informatics, but pointers are most welcome. I will be pursuing this line of research further in other articles at least over the next couple of years.
[Update: presentation now available here.]
Monday, 5 November 2012
Phantoms, RoboCops and teleportation law - just an ordinary day at an extraordinary conference
About a month ago (oh dear) I had again the pleasure of attending GikII, the world's top most number one conference on geek law, this time at the London campus of the University of East Anglia. I also had a presentation of my own there, with the title "Is Botox® the New Tinfoil Hat? On Mind-Reading, Behavioural Biometrics, and Privacy", with the following abstract:
As an example case I used the AVATAR system in pilot use on the US-Mexican border in Nogales, Arizona, only since a couple of months ago. You can find more info on AVATAR here and here. I do feel the need to point out that I did not want to talk specifically about AVATAR, but about the wider privacy implications of face recognition and other biometric technologies in the long term. And to have a bit of fun while doing it, of course.
Somehow I did however manage to briefly mention an issue I seem to return to in every paper, namely the question of judicial (or in this case administrative) decision support. In the EU, the most general regulation of this issue is in the Data Protection Directive (95/46/EC), Article 15, on automated decisions, which can only be warranted by statute or for the purposes of fulfilling a contract. In either case, the system making automated decision must have a safeguard in the form of human supervision with the possibility to override. A similar regulatory scheme is proposed to continue under the forthcoming Data Protection Regulation, this time in Article 20 titled Profiling.
The requirement for human supervision is of course good and necessary, but it is by no means enough by itself. If a system makes correct judgments 90% of the time, people seem to have the tendency to infer that it is correct the other 10% of the time as well. One absurd example of this kind of uncritical sticking to procedure (without machines!) is the Twitter joke trial, which was also taken up by Ray Corrigan in his GikII presentation.
There are different ways to mitigate this. The obvious one is that, especially in a context where the system's decisions are routinely followed, the error rate could not possibly be allowed to be anywhere near 10%. To be sure, rigorous testing protocols are required. One possibility is also to open up the algorithms for review, which can be done ex ante as a part of an authorization protocol, or, especially in an individual case, ex post, or both.
Still, with Big Dada, neither of these is enough of an answer. Creating rigorous tests for real-life systems of this type is easier said than done, and “cheating” on the test by making sure at least all the known test cases work as they should is only common sense. Carving specific requirements for testing in stone is a surefire way to kill all innovation in this field.
Releasing the algorithms isn't panacea, either. When the systems are developed by commercial companies (ahem), there is considerable reluctance (or at least a hefty price tag) for this kind of openness. In a national security setting it Just Isn't Done. And in any case, the sheer complexity of the task means that access to the algorithm is of no use whatever when you are trying to board a flight but THE COMPUTER SAYS NO.
So what's my answer? Quite simple: require that decision support systems are always constructed to give explicit and, upon request, detailed reasons for their decisions in human-compatible terms, just like in a well-written court decision. This makes it easy for anyone to see if there is something completely off in the inputs or the line of reasoning, and step in and override. It also allows for a more qualitative type of testing, when not only the decision but also its rationale can be included in the evaluation. And when the system does something harmless enough (say evaluates likelihood of confusion for trademarks (smiley)), the rationales can be used for educational purposes.
So is this a case for more regulation? Even if it were, the legislator's track record in this field does not exactly promise any immediate relief. One way to solve this is to let the markets decide, but that requires an educated customer base who knows what to require and why. I guess we'll just have to wait and see.
Biometrics are an old acquaintance for data protection law. However, the legal interest has thus far focused on the use of biometrics for identification purposes only, that is, using them as an unique key giving an individual access to something or tying that individual to other, non-biometric personal data. The role of biometric data as (potentially even sensitive) personal data in its own right has yet to received the same kind of attention. Biometrics can also be used for example for personalized outdoor advertising even without positively identifying its individual target. The technological development is extremely fast, and, as with many emerging technologies, law struggles with keeping up to date.Slides here.
One particularly interesting development is the combination of behavioural biometrics with face recognition. Continuous analysis of facial microexpressions based on the Facial Action Coding System (probably most familiar from the TV drama Lie to Me) is being developed for a number of purposes, such as profiling airline passengers and lie detection. For lie detection and `mind-reading' in general, simple optically based biometrics are at least as reliable (ie. not very, at least at this point) as the more widely known fMRI-based and other neuroimaging methods, while being totally noninvasive and thus easy to deploy without the consent or even knowledge of the data subject, and at a fraction of the cost. This type of use of fMRIs and neuroscience in general is already a hot topic in law, but the same questions should be understood more broadly and without commitment to any specific type or level of analysis or any particular technology. Looking for explanations on the neuronal level just confuses the non-specialist completely, thus lending neuroscience explanations their seductive allure.
And so this season’s fashion tip for all paranoiacs is to swap your tinfoil hat for Botox®, as it paralyses the muscles causing facial microexpressions, thus making the technology unreliable. Anti-facial-recognition makeup, which confuses the system by making specific parts of the face undistinguishable, is of course another possibility.
As an example case I used the AVATAR system in pilot use on the US-Mexican border in Nogales, Arizona, only since a couple of months ago. You can find more info on AVATAR here and here. I do feel the need to point out that I did not want to talk specifically about AVATAR, but about the wider privacy implications of face recognition and other biometric technologies in the long term. And to have a bit of fun while doing it, of course.
Somehow I did however manage to briefly mention an issue I seem to return to in every paper, namely the question of judicial (or in this case administrative) decision support. In the EU, the most general regulation of this issue is in the Data Protection Directive (95/46/EC), Article 15, on automated decisions, which can only be warranted by statute or for the purposes of fulfilling a contract. In either case, the system making automated decision must have a safeguard in the form of human supervision with the possibility to override. A similar regulatory scheme is proposed to continue under the forthcoming Data Protection Regulation, this time in Article 20 titled Profiling.
The requirement for human supervision is of course good and necessary, but it is by no means enough by itself. If a system makes correct judgments 90% of the time, people seem to have the tendency to infer that it is correct the other 10% of the time as well. One absurd example of this kind of uncritical sticking to procedure (without machines!) is the Twitter joke trial, which was also taken up by Ray Corrigan in his GikII presentation.
There are different ways to mitigate this. The obvious one is that, especially in a context where the system's decisions are routinely followed, the error rate could not possibly be allowed to be anywhere near 10%. To be sure, rigorous testing protocols are required. One possibility is also to open up the algorithms for review, which can be done ex ante as a part of an authorization protocol, or, especially in an individual case, ex post, or both.
Still, with Big Dada, neither of these is enough of an answer. Creating rigorous tests for real-life systems of this type is easier said than done, and “cheating” on the test by making sure at least all the known test cases work as they should is only common sense. Carving specific requirements for testing in stone is a surefire way to kill all innovation in this field.
Releasing the algorithms isn't panacea, either. When the systems are developed by commercial companies (ahem), there is considerable reluctance (or at least a hefty price tag) for this kind of openness. In a national security setting it Just Isn't Done. And in any case, the sheer complexity of the task means that access to the algorithm is of no use whatever when you are trying to board a flight but THE COMPUTER SAYS NO.
So what's my answer? Quite simple: require that decision support systems are always constructed to give explicit and, upon request, detailed reasons for their decisions in human-compatible terms, just like in a well-written court decision. This makes it easy for anyone to see if there is something completely off in the inputs or the line of reasoning, and step in and override. It also allows for a more qualitative type of testing, when not only the decision but also its rationale can be included in the evaluation. And when the system does something harmless enough (say evaluates likelihood of confusion for trademarks (smiley)), the rationales can be used for educational purposes.
So is this a case for more regulation? Even if it were, the legislator's track record in this field does not exactly promise any immediate relief. One way to solve this is to let the markets decide, but that requires an educated customer base who knows what to require and why. I guess we'll just have to wait and see.
Guest post at VoxPopuLII
I had a guest post on my research in general on the VoxPopuLII blog of the Legal Information Institute at Cornell Law a couple of weeks ago.
Subscribe to:
Posts (Atom)
