AI a ‘game changer’ for document production in construction arbitration

The potential of AI to combat uncontrolled document production was one issue we explored with Lindy Patterson KC, full-time independent arbitrator, dispute board member and adjudicator with 39 Essex Chambers, at a recent event hosted by Pinsent Masons that examined the actions parties to construction arbitration might consider to help make proceedings more efficient.

The balance to strike in controlling document production

Document control seeks to balance two countervailing interests: minimisation of time and costs arising from document production; and ensuring that all the parties have the essential documents required to prove their case or disprove the other party’s case.

In certain cases, as recognised in the 2018 ICC guide, document production may be essential if one of the parties has sole possession of documents needed by the other party. An overly narrow scope of document production may potentially lead to the failure of arbitration as a fair forum to resolve construction related claims, especially those of a more inquisitorial nature, for example, wrongful termination claims, and claims that a bond call is fraudulent or unconscionable.

The 2018 ICC Guide explains the options for approaching document production in arbitration, ranging from no document production, production limited to specific documents or narrow categories of documents which are relevant and material to deciding an issue in arbitration, through to general document production as used in some more adversarial legal systems.

Within this spectrum, recent ‘soft law’ guides consistently recommend limiting document production.

Despite differences in form among the approaches of the soft law guides towards document production control, there appears to be broad agreement that the ambit of appropriate document production is best defined based on the relevant and decisive issues in dispute. This consensus allows for a moderation of document production but not at the expense of denying a party timely access to documents required to prove its case which are not within that party’s possession if the relevant issues so require.

Addressing barriers to controlling document production

There are several issues that present a challenge to controlling document production in construction arbitration. These include the sheer complexity of construction arbitrations, the document-heavy nature of construction projects and therefore construction disputes, and very often the under-developed nature of the parties’ positions at the start of arbitration.

Construction disputes are often complex and document heavy – as matters currently stand there is no real way of getting away from that – but parties can do something about the latter point.

If the parties’ cases are not focussed properly at the outset of an arbitration, there is significant room for ambiguity and therefore a perceived need for counsel on the other side to seek recovery of all documents that might in some way be relevant – because they have no way of knowing which way the case might develop.

Parties should be encouraged to spend the time, and money, at the outset to properly develop their claims and counterclaims so that the other side has a clear idea of the case they are responding to. In our experience, the cases where document production has been more limited have been those where the claimant has properly considered and explained its claims at the outset.

The challenge with this approach is that it needs client buy-in and front-loading of costs, which is not always welcomed. It is also difficult to do when, as happens in a significant number of cases, arbitration is commenced in a rush in an effort to push for an offer of settlement – when that offer doesn’t come, the claimant can find itself in arbitration with limited time to properly bring its case together.

Document production in that sort of scenario is more likely to lead to indiscriminate requests, with enormous numbers of documents being produced. This is problematic for everyone because typically there is not enough time factored into the timetable for disclosure on that scale or for review of all that documentation.

Another issue in controlling document production is how the frameworks for doing so are often interpreted and applied by parties and arbitrators.

Pre-pandemic, it was common to attend hearings with hundreds of files of hard copy documents. Many of these would be common documents: very few documents in the possession of a party would actually move the dial on the issues in dispute. In theory, it is these documents that the document production process should be aimed at getting to.

The test in the IBA Rules on the Taking of Evidence in International Arbitration isn’t simply one of relevance, but rather of materiality to the outcome. All stakeholders in construction arbitration should be focused on ensuring the documents being produced are likely to be determinative of the key issues in the case.

The role for AI

The impact of AI in controlling document production also needs to be considered.

There are solutions becoming available on the market, largely developed in the US, that are aimed at getting the key documents very quickly via machine learning. This can assist not only in reviewing documents as part of the production process, but also in ensuring that parties identify the key documents in their own possession and making sure that their case is properly developed early in the process.

These AI tools can also help in analysing vast quantities of project documents and those provided by the other side. This has the potential to be a real game changer in terms of the time and cost spent on the document collation and review process as a whole.

Cost and time savings in this area should give those involved more leeway to focus in ensuring a fair document production process, and not to sacrifice fairness for perceived efficiency.