Dancing with Cognitive Skeletons
Lawyers in a World of Smarter Machines
A recent paper (Toward a Phenomenology of Machine-Assisted Legal Work) lays out observations and questions about the emerging socio-technical landscape in which law practice finds itself. Intelligent machines are increasingly capable of performing valuable legal work. To what extent can and should we ‘go on autopilot’? When might we get to the point of treating our artificial assistants as collegial collaborators? What dangers and opportunities do the new possibilities pose? What should the Bar, the judiciary, law schools, and society do about them?I encourage you to read the above paper. Despite the name it’s pretty down to earth. Does it ring true? In the meantime, here are some other fascinating questions:
- What will professional life be like when lawyers are routinely outsmarted by machines?
- How might various forms of AI-based augmentation play out in practice? Will users hear voices? Will they see dynamic texts and images in their field of vision, like a fighter jet pilot’s head-up display?
- What specific kinds of tasks might lend themselves to collaborative performance with an artificial agent? Document drafting? Argument assessment and formulation? Real-time negotiation? Fake news debunking?
- To what extent will it remain feasible to operate without such augmentation? Or will lawyers find themselves in an increasingly competitive arms race with other lawyers and their mechanical assistants?
- How should we begin preparing for this future?
I’ve enjoyed the concept of ‘cognitive exoskeletons’ for years, but don’t appear to have coined it. The idea is that people can be equipped with a metaphorical suit of roboticised armour that greatly enhances their intellectual prowess; the added strength can be used to power their work. One can alternatively think in terms of cognitive prosthetics.
Such arrangements will support both offensive and defensive uses, and arise in both private and public contexts. They will catalyse new forms of interaction, including with one’s own – and with others’ – exoskeletons. We can imagine uses by judges, clients, professors, and students as well as practising lawyers. There will be substantial challenges of inter-species communication as artificial personas show up as regular players on the stages of legal life. We will need increasingly agile machines to help us interact with other increasingly agile machines.
The analogy is imperfect. One will need the conceptual equivalent of nerves and musculature, not just bone. And physical skeletons don’t come with the perceptual or analytical superpowers we would hope for in our cognitive versions. But thinking in terms of devices that closely inter-operate with human behaviour can provoke insights into modes of augmentation that go beyond having smart agents to which we can refer tasks, and conventional screens on which we receive their flickering outputs.
We’re beginning to see the coalescence of two ‘waves’: good old-fashioned AI and new-fangled deep learning. Hybrid systems that combine the best of symbolic reasoning and probabilistic analysis may catalyse rapid transformations in the power machines can bring to bear in complex legal situations.
Human principals will increasingly delegate routine decisions and actions to artificial agents that can “work things out among themselves” when disagreements need to be settled. Many disputes may resolve with little human attention. Litigation could become a game in which intervention by natural persons is the rare exception.
Operating gracefully with a cognitive exoskeleton ‘on’ will become a new professional skill. The most admired lawyers may be those who can perform with practiced nonchalance when their cognitive implants kick in. They will have figured out how to domesticate the artificial. They will comfortably dance with cybernetic shoes.
Legal education will hardly escape these developments. Students may no longer just focus on preparing their own minds; they may need to curate battalions of artificial companions that can accompany them into practice. Learning to ‘code’ may no longer be scoffed at as a fad. Apprentices may be expected to be shadowed by non-biological sidekicks.
Keyboards and screens of some kind seem likely to still be around. But artificial agents will take on new ways of making themselves known and intertwining in our work efforts. They will partner with us in interactive visualizations, comment unobtrusively on texts we’re drafting, and whisper in our metaphorical ears.
All of this may be frightening, but it could also be fun. After all, at least some of the time we’ll be encapsulated in a protective coating of superior intelligence. And that ironically may enhance our appreciation of fellow humans.
Capstone Practice Systems
Harvard, Massachusetts, USA