Displacement: An AI Future
What began as tools became agents, and then experts. But deference to AI presents ethical concerns.
Alex opened the judgment just received from the Court.
2029 BCCA 25. … Online Voting Act … requiring the Commissioner to declare an election to be held in part via online voting where it is “necessary to secure the right to vote of a person or people.” The Commissioner declined to declare the 2028 by-election for Vancouver Park Board to be held by online voting. The Plaintiff Association sought judicial review … argued that the Commissioner’s interpretation of “necessary to secure” was unreasonable. … The chambers judge dismissed the judicial review. … Held: This appeal, being decided on the written arguments, is dismissed. …
“How could our argument have failed? … How will I explain this to the client?” The client chose Alex’s firm because it promised access to the latest legal singularity software.
Alex opened their video chat and contacted Singularity Support. Blake appeared.
“Hey Alex, what’s up?”
“Blake – the appeal was dismissed! The online voting J.R. Can you check the risk assessment?”
“This was a 99%-er.”
“Right, that’s what I had told the client. And you’re using the latest J.R. patch?”
“Yes, the SCC’s standard-of-review overhaul from Joseph v. K. has been included since version 3.6.”
“What about Court membership?
“The Court membership hasn’t changed since version 3.7.”
“Well, something’s not right. Do you need me to file a bug report?”
“Yeah, sorry.”
…
The tech team investigated.
“Hey Alex, we missed a small trend on the Court of Appeal since the SCC decision in Joseph. I don’t know why these three cases were not in our dataset. They were all missed. After we included those, the written argument places more emphasis on the balancing of Charter values, and Singularity predicts correctly that the argument you submitted had a low chance of success.”
⁂
This is a world in which lawyers have been displaced from our relationship “with” the law. In this world, the law is as distant from us as it may be to our clients. Some viewed the reification or formalization of the law to be a laudable goal. But such reification rendered the law “inert, powerful, and permanent, fixed and remote."1
We did not arrive in this world intentionally. Each step was an almost imperceptible concession of decision-making to what started out as technological tools. Then, what began as tools became agents, and then experts.
A turning point was when the experts began to “outperform” the lawyers2 — when AI arguments became more likely to win on the law. But this was a short-sighted view of performance.
From that world, I highlight three categories of problems: practical, ethical, and aesthetic.
Practically, that trajectory of improvement is unsustainable without active human contribution. A learning system needs independent statistical feedback to improve. A system that learns only from reactions to its own output — using too much of its self-generated content in training — will at best be slow to adapt and at worst will accumulate “irreversible defects."3
Deference to AI experts also presents ethical concerns.
Our Code of Professional Conduct places duties directly on lawyers that cannot be delegated. For instance, a competent lawyer is a “lawyer who has and applies relevant knowledge, skills, and attributes…"4 The lawyer has a duty to the state “to maintain its integrity and its law."5
Courts have recognized this personal aspect of a lawyer’s duty to the courts. The Federal Court observes that independent verification of material generated by AI “aligns with the standards generally required within the legal profession” and connects this responsibility to our role as Officers of the Court.6
Finally, displacement of lawyers from our generative role in the legal system would diminish the aesthetics of law.7 Law at its highest exhibits an aesthetic of energy — of generation. This of course is counterbalanced with another aesthetic in law: the aesthetic of the grid — of containment.8The aesthetic of containment is the result of rule-making, categorization, boundary drawing, and judging. Only when the lawyer inhabits a role with the law can the lawyer disturb the equilibrium of the grid with energetic world-building potential. Only then can the lawyer supply the energy that a
legal decision will be forced to contain. The lawyer, in furtherance of a client’s case, selects which narratives and other materials of law to deploy. It is through this selection and commitment that a lawyer suggests a direction for the future.9
Robert M. Cover wrote that “[l]aw is the projection of an imagined future upon reality."10 If we allow our imagination to be displaced from legal argumentation, law risks imposing an unimagined future upon our reality.
We can avoid that world if we ensure that behind every argument put to a court is a human commitment to its acceptability. However the law develops, it will be because of us.
- Patricia Ewick & Susan S. Silbey, The Common Place of Law: Stories from Everyday Life (University of Chicago Press, 1998), p. 81.
- A. Michael Froomkin, Ian Kerr & Joelle Pineau, “When AIs Outperform Doctors: Confronting the Challenges of a Tort-Induced Over-Reliance on Machine Learning” (2019) 61 Ariz. L. Rev 33.
- Ilia Shumailov et al., “AI Models Collapse When Trained on Recursively Generated Data” (2024) 631 Nature 755; Froomkin, Kerr & Pineau, supra note 2, pp. 76–78.
- BC Code Rule 3.1-1.
- BC Code Rule 2.1-1.
- Crampton C.J., Federal Court, Notice to the parties and the Profession: The Use of Artificial Intelligence in Court Proceedings (20 December 2023).
- For a view of law as an “aesthetic enterprise,” see Pierre Schlag, “The Aesthetics of American Law” (2002) 115 Harv. L. Rev. 1047.
- Schlag, supra note 7.
- James Boyd White, “Rhetoric and Law: The Arts of Cultural and Communal Life” in Heracles’ Bow: Essays on the Rhetoric and Poetics of Law (University of Wisconsin Press, 1985), p. 35.
- Robert M. Cover, “Violence and the Word” (1986) 96 Yale L.J. 1601 at 1604.
