Law as Code: a legal system shaped by software – a response

Joshua Browder is a provocative, thoughtful and successful entrepreneur. His DoNotPay empire , describing itself as ‘the world’s first robot lawyer’, expands by the day. And he has now in his own words ‘finally published my writing on the future of the legal system’. Mr Browder was raised in Britain but did not get where he has through British understatement. His summary for twitter is pretty straightforward: ‘Expensive lawyers will be replaced by software. Here’s how it will happen.’

Mr Browder’s article is actually a bit more nuanced. It is worth reading. Hopefully, it will contribute to debate on how technology might – and should – change the law. This is my summary of his argument. put as neutrally as I can. The cost of lawyers is close to prohibitive for many people and businesses. However, the current system makes them unavoidable: ‘people who hired an attorney to help them with disability benefits cases were three times more likely to be successful than people who did not.’ Lawyers have a long history of maximising and monetising their role. This goes back to our Edward 1 at least. [So far, surely so uncontroversial.]

We should use software to ‘scale the expertise of a limited number of lawyers’. He then gives a really interesting example. Some forms, eg for refugees, may ask if someone is fluent in English. For a variety of reasons, the user may say that are when they are not. Software would allow the incorporation of warnings, tests, examples that would help someone to show their level of language proficiency and avoid problems that might otherwise arise from its overstatement. This is a good point. It could be extended to software prompts that encourage users to give information in support of a claim or position: ‘software can not only save time … but can replicate the “insider expertise” of a lawyer explaining a form a number of times’. This would, of course, depend on a government that wanted claimants to maximise their chances of success. Mine cannot be the only jurisdiction where this involves somewhat of a leap of faith.

His second major point is that all legislation should be available free on the net. In Georgia apparently, a full set of laws costs $1207.02. Well, it shouldn’t. Mr Browder is right. Laws are public: that is their point. They should be freely available to those affected by them. But, the argument gets taken forward. If laws were code then they could be more easily amended and this would ‘allow the public to assist officials in discovering loopholes and special interests’. Yes. Well that would be fine while they were in draft and unpassed – which, to be fair, is precisely Mr Browder’s argument. But it would be a bit chaotic if this kind of amendment was allowed post-implementation. And, at least in the UK, you can get the drafts from the Parliament website. You just need to be an expert to know what they mean.

The third argument is that ‘software applied to legal code would … represent the building blocks to automate legal decisions, skipping the court room entirely … undisputed legal cases lend themselves well to this sort of automation’. This is correct and, indeed, the basis of the DoNotPay modus operandi. It is also the increasing basis of much decision-making by government. Fail to pay the congestion charge in London or the toll to cross one of the bridges over the Thames and you get automated decision-making with the additional plus of discretionary waiver. Fine.

But how far can this extend? Mr Browder muses: ’Over time, such automatic decisions can be extended to the vast array of uncontested disputes that involve jumping through bureaucratic hoops.’ Alas, one person’s hoop is another’s shield. His examples of prime candidates for automation in due course are ‘divorces, planning applications, traffic appeals and my other processes’. Well, the UK may be different but planning applications are a pretty contested area here. Divorce – if you include custody and maintenance – is pretty contested everywhere. 

His conclusion: ‘a software-first approach does not have to be all or nothing … while tech is unlikely to completely overtake justice system, automating minor or uncontested case could free up the (indisputably backlogged existing system … By embracing software and dismantling information silos, we can reshape the legal process … [towards] a system that truly works for all’. Surely an impeccable – indeed, nuanced – conclusion.

There may be an issue of culture here. Is the best way to get change to announce a big headline and hide the detail in the small print. Or do you get better results from less ambitious, limited claims and advances? It seems to me that Mr Browder’s argument is not really that the status of law should change so that we see it as software but rather that we seek to maximise how software might be used to maximise its use by all citizens at minimum cost. If that is the case then the following propositions must be right:

  1. There should be no cost or copyright restrictions on the use of laws- statute of caselaw.
  2. Governments should build, after consultation with all interested parties, automated ways in which users can maximise their understanding of laws that apply to them and any case that they might want to make. This is unlikely to be via statute itself but it could be ‘bolt on’ additions like BC’s Civil Resolution Tribunal Solution Explorer. Want people to show they are fluent in English, build in ways of their providing it. Good point.
  3. Decision-making should be automated where there is minimal discretion – though with non-digital access to those otherwise excluded and appeal rights that allow human supervision of automated processes.
  4. The limiting effect of excessive cost to access to justice cannot be swept under the carpet and urgently needs addressing. Technology should help.

Technology is unlikely, if you implemented it properly, to lead to the replacement of expensive lawyers by cheap sofware. They actually have less to fear than politicians who might find obscurity and confusion politically helpful.  What it would do is lead to displacement in lawyers’ activity. So, if governments really want those affected to be able to claim Windrush compensation and disability benefits (two British examples of current government obfuscation, delay and resistance) then they should apply software to the law to help them. And we can increasingly see how that could be done.

Just don’t hold your breath.

And well done, Mr Browder, for keeping the argument going even at the risk of a bit of overselling. Now we need to hear from the politicians.

Leave a Reply