Stuff AI CAN'T Do

¿Puede la IA leer un contrato y sentir dónde está la trampa ?

¿Qué opinas?

Los abogados ganan sus honorarios con esto. La cláusula que parece correcta pero en la práctica significa algo diferente en esta jurisdicción con esta contraparte.

Background

Lawyers are compensated for spotting contractual ambiguities that appear innocuous but carry significant implications in specific jurisdictions or with particular counterparties. Current AI systems excel at clause extraction, risk flagging, and term comparison by processing large datasets, yet they face limitations in contextual comprehension and subjective judgment. AI highlights potential ambiguities or clashes with standard templates but often lacks the ability to capture the full complexity of human language or infer unstated consequences. Scholars at Stanford Law School (enriched May 9, 2026) emphasize that while AI can automate routine review tasks—such as identifying mismatches with predefined rules—it cannot yet replicate human intuition or contextual awareness when detecting traps like hidden liabilities or misaligned obligations. As of May 11, 2026, research continues to focus on advancing AI’s interpretive depth, though the identification of subtle contractual pitfalls remains primarily within the purview of legal professionals leveraging both analytical tools and experiential insight.

Estado verificado por última vez en May 14, 2026.

📰

Galería

In the Court of AI Capability
Summary of Findings
Verdict over time
May 2026May 2026
Sitting at the Bench Filed · may. 14, 2026
— The Question Before the Court —

¿Puede la IA leer un contrato y sentir dónde está la trampa?

★ The Court Finds ★
▲ Upgraded from No
Casi

Existen demostraciones limitadas — pero el panel no fue unánime.

Ruling of the Bench

The jury found AI capable of mapping the obvious tripwires in a contract—count the caps, tally the ticks, flag the outlier phrasing—but concluded it still wobbles in the dark corners where language bends human meaning. A three-to-one split on “almost there” showed unanimous agreement that the tools are useful, yet unanimous hesitation to trust them alone. The Court rules: AI can spot the trap, but it still stumbles in the shadow.

— Hon. E. Dijkstra-Patel, Presiding
Jury Tally
0
4Casi
0No
Verdict Confidence
78%
The Court of AI Capability is, of course, not a real court.
But the data is real.
The Case File · Stacked History
Session I · May 2026 No
Case № 0337 · Session II
In the Court of AI Capability

The Case File

Docket № 0337 · Session II · Vol. II
I. Particulars of the Case
Question put to the court¿Puede la IA leer un contrato y sentir dónde está la trampa?
SessionII (2 hearing)
Convened14 may. 2026
Previously ruledNO (May '26) → ALMOST (May '26)
Presiding JudgeHon. E. Dijkstra-Patel
II. Cumulative Tally Across Sessions

Across 2 sessions, 7 jurors have heard this case. Combined tally: 0 YES · 4 ALMOST · 3 NO · 0 IN RESEARCH.

Note: cumulative includes older juror opinions. The current session tally above is the live verdict.

III. Verdict

By a vote of 0 — 4 — 0, the panel returns a verdict of CASI, with verdict confidence of 78%. The court so orders. Verdict upgraded from prior session.

IV. Declaraciones del tribunal
Jurado I ALMOST

"AI can analyze contracts for potential issues"

Jurado II ALMOST

"AI can identify some traps in plain-language contracts with partial reliability but struggles with nuanced legal language or buried clauses."

Jurado III ALMOST

"AI can detect anomalous or high-risk clauses in contracts using pattern recognition, but cannot reliably 'feel' subjective traps without explicit training on adversarial examples."

Jurado IV ALMOST

"AI can analyze contracts but may miss nuanced traps"

Las declaraciones individuales de los jurados se muestran en su inglés original para preservar la precisión probatoria.

E. Dijkstra-Patel
Presiding Judge
M. Lovelace
Clerk of the Court

Lo que el público piensa

No 59% · Sí 10% · Quizás 31% 164 votes
No · 59%
Quizás · 31%
La tendencia necesita votos de al menos 2 días distintos.

Discusión

no comments

Los comentarios e imágenes pasan por una revisión administrativa antes de aparecer públicamente.

2 jury checks · más reciente hace 13 horas
14 May 2026 4 jurors · indeciso, indeciso, indeciso, indeciso indeciso estado cambiado
12 May 2026 3 jurors · no puede, no puede, no puede no puede

Cada fila es una comprobación de jurado independiente. Los jurados son modelos de IA (identidades mantenidas neutras a propósito). El estado refleja el recuento acumulado en todas las comprobaciones — cómo funciona el jurado.

Más en Judgment

¿Nos faltó uno?

Revisamos semanalmente.