eDiscovery Leaders Live: Matthew Gobal of Gilbert + Tobin

    George Socha
    George Socha

    eDiscovery Leaders Live: Matthew Gobal of Gilbert + Tobin


    Matthew Gobal, Director - Legal Informatics and R+D at Gilbert + Tobin, joins George Socha, Senior Vice President of Brand Awareness at Reveal, for ACEDS #eDiscoveryLeadersLive.

    Matthew leads a specialized in-house multidisciplinary legal informatics team at Gilbert + Tobin that uses a variety of data analytics, eDiscovery, and other AI technology tools to facilitate large scale document/data production and review for commercial litigation, regulatory compliance (ATO, ASIC, ACCC) and due diligence for mergers and acquisition transactions. Matthew has more than 20 years of experience in the legal technology industry, including with two of Australia's preeminent law firms, Allens for 13 years and Gilbert + Tobin for the past 11 years. Matthew was an inaugural recipient of the International Legal Technology Association's (ITLA's) Browning Marean Scholarship in 2015, providing international recognition of his leadership in eDiscovery and litigation technology.

    Matthew shared his thoughts on generative AI, the status of eDiscovery in Australia, and the work done by the EDRM DupeID project.

    Key Highlights

      • [1:06] Introducing Matthew.
      • [1:18] Matthew’s background.
      • [4:41] What brought Matthew to discovery in the 1990s and then eDiscovery.
      • [7:44] eDiscovery is to libraries as….
      • [8:59] Where you can with today’s technology.
      • [10:49] His current role and team: a center of excellence for eDiscovery and data.
      • [11:03] The potential of generative AI.
      • [11:43] G+T’s focus on generative AI.
      • [12:31] The limits of how to use generative AI with highly sensitive client data.
      • [13:00] The awkward side of generative AI: hallucinations.
      • [14:52] Australian law firms’ relatively high level of technological competence.
      • [15:57] Australian Federal Court’s well-established eDiscovery protocol.
      • [16:24] The many significant eDiscovery things that have happened in Australia.
      • [16:58] Normalizing the eDiscovery process against a single Federal standard.
      • [18:24] Today’s eDiscovery headaches: other data models, cloud, encryption, mobile, chat, recordings.
      • [19:13] Perhaps the biggest eDiscovery challenge in Australia: the whole model is based on a “document”, and we no longer live in that world.
      • [19:50] Big Australian law firms’ focus on technical capabilities rather than large teams.
      • [21:04] The value of Australian pragmatism.
      • [22:10] The EDRM DupeID project: what it is, what it offers, and why it matters.

    Key Quotes 

    • “With today’s predictive coding, machine learning models, data categorization, classification, where you can look at trends and statistical sampling – these are amazing ways with, compared to the linear effort, fairly minimal effort to get enormous return in having … certainly over how far you will be reviewing, how much that review will cost, and what you’re likely to find in that review….”
    • “There’s enormous potential I think to transform all business and all corporate life with, as a positive, in harnessing generative AI and working out efficiency in finally having technology do away with mundane administrative low level, low value work, drudgery, that unfortunately clogs up all of our days.”
    • “I don’t think we’re yet at a point where, we all [have] gotten comfortable with how do we engage with generative AI with highly sensitive client data.”
    • “These are general-purpose models trained on enormous data sets of general-purpose content. The system is simply predicting a pattern based on the inputs you’ve given it … and then matching that against what it knows from its training. If you don’t have a training corpus that is tuned to your specific niche, then you’re going to get sometimes fairly anomalous, as in extremely frightening, results.”
    • “If you don’t have enough knowledge to verify the answer [given by generative AI], it’s extremely dangerous. If you don’t know the answer is wrong – if it’s completely made-up case law or legislation, you’ve got a major issue in verifying that.”
    • “I suspect the Australian market as a whole has a fairly high level of competence and education and proficiency in the use of technology.”

    Connect with Matthew