E-discovery

The Devil’s Advocate™ →Go uptown Discovery before the information revolution In which the curmudgeonly old sod puts the world to rights. Index — Click ᐅ to expand: The Devil’s Advocate Tell me more Sign up for our newsletter — or just get in touch: for ½ a weekly 🍺 you get to consult JC. Ask about it here.

The JC rarely strays into matters of litigation, regarding court-based resolution of commercial disputes as a kind of admission of liability in itself.

But civil procedure is an interesting prism for a (literally) disinterested observer to view the promise — or existential threat — of legaltech. For while there are as many reasons for la révolution tardive as there are thought leaders to formulate them — more, in fact — all are as one when reaching for examples of how legal technology has already changed things for the better: e-Discovery is the anecdote of choice.

Once upon a time, before the internet was invented, the JC was a clerk in a commercial litigation firm in the antipodes. One of the most tedious jobs was preparing lists of documents for discovery — disclosure to the opposing side — in forthcoming litigation. There could be hundreds of relevant documents. Even thousands. The poor clerks had to identify each, name it, date it, record by whom it was sent and to whom, and designate its privilege.

Now this being before email — the fax machines was à la mode — there weren’t that many documents, nor any automated means of cataloguing them.

How times have changed. The information revolution has created not just a new class of electronic documents with new classes of metadata but a whole new volume of information: the ease of creating, changing, duplicating and circulating means that the number of potentially relevant documents is as likely to be be in the millions as the thousands.

Automated tools, using neural networks, machine learning and other kinds of AI have emerged to handle discovery analytics.

Parkinson’s law for the digital age?

What follows is a needlessly formal argument for the proposition “the cost a person will incur to resolve a dispute is a function of the amount at stake and the merit of her claim and not the tools available to conduct the litigation”. In other words, benefits of legaltech accrue to practitioners and not customers.

Those selling e-discovery platforms doubtless are thrilled at this technical revolution. Perhaps, too, court lawyers — though there is an element of an arms race about it. But the question we pose is a different one: how has legaltech in general, and e-discovery in particular, contributed to legal outcomes for clients?

E-discovery exists to solve a problem that was only created by technology in the first place. As we have remarked elsewhere, before the advent of email — before lawyers had screens on desktops and were dependent on paying someone to put their words onto paper, the very act of generating a document, let alone sending one, was orders of magnitude more difficult and expensive. This provided strong incentives not to create unnecessary documents and to make sure the documents one did create were short and to the point.

These incentives no longer remain. The volume of information that now passes is monstrous; its quality is dismal. To the extent it is the only means of handling this volume, this is not e-discovery solving a problem. This is technology creating one.

We wonder, too, whether the advent of e-discovery has made litigation cheaper, faster or more effective. Litigation does not, generally, seem to be getting cheaper.

Have commercial disputes become intrinsically more complicated since the information revolution, or is their formal complication a function of how lazy technology has allowed us to become? There are ten times as many documents, but is there any more signal, or is it just a lot more noise? Is e-discovery really just an articulation of Parkinson’s law for the digital natives?

A dispute has, from a plaintiff’s perspective, a reductively simple complexion:

“I believe I am entitled to value X, but I have received value Y”.

This is, in the vernacular, a univariate calculation. There is an amount to be claimed — X - Y — and a maximum amount one is prepared to pay to recover that amount — (X - Y)/n — where n will vary depending on the strength of customer’s original conviction, but logically must be greater than 1.

For those engaged to help resolve the dispute, however, the problem is multivariate: how reasonable is your client’s assessment of its own chances; how solvent is the defendant; how likely is it to defend, how vigorously, and how skilfully, which forum is most convenient to resolve that dispute; how sympathetic is that forum likely to be — each of these factors one can assign a probabilistic assessment at best. They are known unknowns. The great skill of the advocate is to assess them: should the assessment come inside the customer’s calculation, there is a livelihood to be made.

The cost of discovery is part of that assessment. Since we can now analyse, tag and redact cheaply, more material is in scope for formalistic arguments.

Now no-one would suggest, God forbid, that lawyers gratuitously run down blind alleys to clock fees against their own clients, but it is a well-established technique to do this against the opponent’s client, precisely to present the cost of ongoing litigation as an incentive to early settlement.

We hypothesise, without any evidence whatsoever, that the number and type of documents having a measurable impact on the outcome of litigation is broadly the same in 2022 as it was in 1980. We doubt many documents of this nature would have been turned up by e-discovery but not manual discovery. That is to say, one would find all smoking guns (needles?) in the part of haystack one could search by hand, and not the great remainder that can only be covered by machines. If that is the case then we wonder whether e-discovery has made any difference to customer outcomes at all.

See also

• Innovation paradox
• Queen’s Counsel
• Writing for a judge
• Parkinson’s law