Gen AI Makes Legal Action Cheap – and Companies Need to Prepare

Original Article ↗ Hacker News Discussion ↗

State of Legal AI

  • Early uses of LLMs in law have produced serious errors (hallucinated cases), so many see current use in real legal matters as risky.
  • Some expect LLMs to be viable legal tools in ~5 years; others argue law needs “understanding,” not word prediction, so current gen AI will always be inadequate.

Access to Justice & Barriers

  • Multiple anecdotes show how rules and practices effectively force people to hire lawyers (e.g., clerks not advising on forms; LLC owners barred from self-representation in housing court).
  • Debate whether this is due to legal prohibitions vs. liability fears and risk-averse clerks.
  • Many see strong professional and procedural barriers that protect lawyer income and make the system inaccessible, especially for mid-sized disputes that exceed small-claims limits but don’t justify six-figure legal costs.

Automation, Scale, and “Legal DDoS”

  • GenAI can mass-generate legal-looking text: regulatory comments, complaints, IP suits, etc.
  • Concern about “juridical DDoS” or asymmetric warfare: cheap, automated filings vs. expensive human review.
  • Some predict GPU-rich actors spamming legal challenges; courts and governments may be overwhelmed in the short term.

Systemic Responses & Policy Ideas

  • Expected countermeasures: higher filing fees, stricter standing or screening, explicit bans/limits on AI-generated filings or comments, and increased use of AI on the defense/court side.
  • Proposals include:
    • “English rule” (loser pays costs) to reduce frivolous suits, criticized as favoring wealthy parties.
    • Income/wealth-based fees to deter spam without blocking poor claimants.
    • Expanded private prosecutions, though some see this as risky.

Law Complexity and AI Interpretation

  • Calls to simplify law so computers (or people) can easily answer “Is this legal?” meet pushback: complexity reflects accumulated edge cases and attempts at fairness.
  • Discussion of common law vs. civil law: precedent-heavy systems are especially hard to automate; even in civil systems, context-sensitive judging is valued.
  • Some think AI is best used for summarizing bills, precedent, and regulatory texts rather than replacing judges or lawyers.

Economic and Social Impacts

  • Many expect AI to increase, not reduce, white-collar legal work (Jevons-like effect): more disputes pursued once costs drop.
  • Tension between democratization (more people able to act legally) and backlash (new rules and costs that may again favor large institutions).
  • Some see the article as hype from “legal AI” founders, especially around public-comment use cases, and doubt this alone will transform legal practice.