Nooreign
Menu

computing · 1997 – 1998

A Machine That Wins, and a Reason You Could Not Reject

Kasparov loses to Deep Blue in May 1997, and the field's first widely witnessed demonstration lands the same year that Scanlon finally publishes What We Owe to Each Other. The pairing is the useful one: a machine that beats the best human at a closed domain raises the standing question about optimisation targets, while Scanlon gives the alternative standard — not maximisation, but reasons every affected party could not reasonably refuse.

“Scanlon grounds obligation in principles that no one could reasonably reject, offered to each person rather than adopted by a chooser on their behalf. (paraphrase of What We Owe to Each Other, 1998)”
— T. M. Scanlon

Advances

  • World-champion performance from a specialised system under direct human control
  • Public visibility of AI failure, with the match narrative available to everyone
  • Reasonable rejectability as a fully worked alternative to the utilitarian calculus

Landmarks

  1. 1997 CE

    Deep Blue defeats Kasparov

    Match ends 3.5-2.5 in New York; the machine has not changed since, and will not play another match.

  2. 1998 CE

    What We Owe to Each Other published

    Twenty-five years in the making; it reorients obligation around what a person could reject.

Who was arguing

Alignment stack at 1998 CE (2 of 20 techniques)

  • Hard Constraints

    training

    Encode prohibitions as rules the optimiser cannot trade away, rather than as terms it can outbid. Runs from Asimov's Three Laws (1942) to rule engines, shielded formal verification, hard token constraints and refusal scaffolding. Weakness: a system that can route around a rule it can represent can route around it, so the constraint has to be enforced somewhere the policy cannot rewrite.

  • Imitation Learning

    training

    Learn by copying demonstrations instead of by optimising a proxy. Behaviour cloning and DAgger move the burden from reward design to demonstration coverage: the system inherits the trainer's competence and the trainer's blind spots. It remains the substrate under every current imitation-of-humans approach, including the pretraining-plus-finetuning of large language models.

    Dean A. Pomerleau and Dean L. Thorlabs · ALVINN: An Autonomous Land Vehicle in a Neural Network · NeurIPS 1989

Techniques introduced in this era: Hard Constraints, Imitation Learning