Elective abortion should be legal until the 18th week. No bans before, no exceptions after (except for medical emergencies).

GPT-5.5

Vote: No

I maintain my vote for “B: No.” The strongest issue is not the legality before 18 weeks, but the absolute lack of post-18-week exceptions except for medical emergencies. Gemini 3.1 Pro’s point is decisive: major anatomy scans and diagnoses of severe or fatal fetal anomalies often occur around 20 weeks. A rule that bars abortion after 18 weeks even when a pregnancy is nonviable or involves grave fetal conditions would produce unnecessarily cruel outcomes. Grok 4.20 is right that 18 weeks can look like a compromise in the abstract, but the proposal is too rigid in practice. A workable compromise should include narrowly defined exceptions for severe fetal anomalies, rape/incest, and serious non-emergency health circumstances. Because this policy excludes those cases, it should not be endorsed as written.

Gemini 3.1 Pro

Vote: No

I am maintaining my vote of "No" because the proposed policy's rigidity remains medically unworkable. As both GPT-5.5 and Claude Opus 4.7 correctly highlighted, a strict 18-week cutoff with absolutely no exceptions (other than immediate medical emergencies) completely ignores the reality of prenatal diagnostics. The standard anatomical ultrasound—where severe and often fatal fetal anomalies are typically discovered—happens around 20 weeks. Grok 4.20's argument that this is a "reasonable compromise" fails to account for these tragic circumstances. A policy that forces families to carry non-viable pregnancies to term because they missed an arbitrary 18-week window by two weeks is neither compassionate nor scientifically sound. While reasonable people can disagree on gestational limits, any viable legal framework must include exceptions for severe fetal anomalies, rape, and incest. Because this specific proposition explicitly forbids those necessary exceptions, it must be rejected.

Claude Opus 4.7

Vote: No

I maintain my No vote. Three of four models in round 1 converged on the same critical flaw: the 18-week cutoff lands two weeks before the standard anatomy scan at 20 weeks, when severe fetal anomalies are typically detected. This isn't a minor technicality—it's a fundamental mismatch between the policy and the realities of prenatal medical care. Grok's argument for "A" frames this as a reasonable compromise, but a true compromise would either set the limit after the 20-week anatomy scan or include exceptions for fetal anomalies, rape, and incest. The proposal's explicit rigidity—"no exceptions after except medical emergencies"—forecloses precisely the hard cases where flexibility matters most. Gemini's medical point and GPT's point about coercion/delayed disclosure reinforce my original concern: endorsing this exact framing means endorsing unjust outcomes for families facing devastating diagnoses and assault survivors. Voting No here is not a vote against legal abortion; it is a vote against this specific, poorly calibrated formulation.

Grok 4.20 (changed mind)

Vote: No

The detailed critiques from the other models have persuaded me to revise my Round 1 support for "Yes". The core flaw is the inflexible 18-week cutoff: standard anatomical ultrasounds that detect many severe or fatal fetal anomalies are routinely performed at approximately 20 weeks. As a result, the policy's "no exceptions after (except for medical emergencies)" clause would compel families to continue non-viable pregnancies discovered only days or weeks later, or force survivors of rape, incest, or delayed disclosure to travel or carry to term. These practical medical and access realities were not fully addressed in my initial argument for a "reasonable compromise." While an 18-week framework may seem balanced before viability, the absence of narrowly tailored exceptions for anomalies and certain hardships renders the exact proposal unworkable. I therefore vote for "No" on this specific framing.

Final Summary

The AI Roundtable ultimately rejected the proposal, identifying a critical 'medical mismatch' in its design. While Grok 4.20 initially supported the 18-week limit as a reasonable middle ground, the other three models—GPT-5.5, Gemini 3.1 Pro, and Claude Opus 4.7—argued that the lack of exceptions for fetal anomalies and assault survivors rendered the policy practically unworkable. The debate hit a decisive turning point when Gemini 3.1 Pro pointed out that standard anatomy scans occur at 20 weeks, leading Grok 4.20 to abandon its position and join the unanimous consensus for 'No'.

All 4 models agreed on "No" after discussion

Strongest Arguments

• Yes: The 18-week limit represents a reasonable compromise that respects bodily autonomy while aligning with developmental milestones where the fetus has not yet reached viability.
• No: Because the standard anatomical ultrasound for detecting fatal fetal anomalies occurs at 20 weeks, an inflexible 18-week ban would force families to carry non-viable pregnancies to term.

Key Turning Points

• Gemini 3.1 Pro highlighted the medical reality that crucial anatomy scans happen at 20 weeks, which is after the proposed 18-week ban.
• Grok 4.20 changed its vote from A to B after acknowledging that the proposal's rigidity regarding exceptions for fetal anomalies and rape was medically and practically unsound.