🔥 Hot topics · KAN IKKE · Kan · § The Court · Seneste omvendinger · 📈 Tidslinje · Spørg · Ledere · 🔥 Hot topics · KAN IKKE · Kan · § The Court · Seneste omvendinger · 📈 Tidslinje · Spørg · Ledere
Stuff AI CAN'T Do

Kan AI identificere hadeful tale på produktionsskala ?

Hvad mener du?

Ufuldkommen, omstridt og konstant genoptrænet — men hver større platform kører et automatiseret lag, der markerer eller fjerner de fleste tilfælde uden menneskelige øjne.

Background

Current AI systems can identify hate speech in text with reasonable accuracy, using machine learning models trained on large datasets of labeled examples (Association for Computational Linguistics, 2026). However, achieving high accuracy at production scale is challenging due to the nuances of language, context, and the evolving nature of hate speech. To address these challenges, researchers and developers are exploring techniques such as transfer learning, ensemble methods, and human-in-the-loop feedback. Imperfect, controversial, and constantly retrained, every major platform runs an automated layer that flags or removes most cases without human eyes. As a result, many social media and online platforms have begun to deploy AI-powered hate speech detection systems to moderate user-generated content.

Status senest tjekket July 3, 2026.

📰

Galleri

In the Court of AI Capability
Summary of Findings
Verdict over time
May 2026May 2026May 2026May 2026May 2026May 2026Jun 2026Jun 2026Jun 2026Jun 2026Jun 2026Jul 2026
Sitting at the Bench Filed · jul. 3, 2026
— The Question Before the Court —

Kan AI identificere hadeful tale på produktionsskala?

★ The Court Finds ★
Reaffirmed
Ja

Juryen fandt et klart bekræftende svar.

Ruling of the Bench

After hearing expert testimony on standardized benchmarks and real-world deployment, the jury unanimously agreed that current AI systems are capable of identifying hate speech at production scale. They credited the strong performance metrics and operational reliability of existing tools, finding no meaningful gap between capability and real-world application. The ruling: "The gavel falls—AI already polices the digital streets.

— Hon. E. Dijkstra-Patel, Presiding
Jury Tally
2Ja
0Næsten
0Nej
Verdict Confidence
93%
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 In_research
Session II · May 2026 Ja
Session III · May 2026 Ja · 79%
Session IV · May 2026 Ja · 84%
Session V · May 2026 Ja · 84%
Session VI · May 2026 Ja · 72%
Session VII · Jun 2026 Ja · 82%
Session VIII · Jun 2026 Ja · 83%
Session IX · Jun 2026 Ja · 93%
Session X · Jun 2026 Næsten · 89%
Session XI · Jun 2026 Ja · 98%
Case № 1101 · Session XII
In the Court of AI Capability

The Case File

Docket № 1101 · Session XII · Vol. XII
I. Particulars of the Case
Question put to the courtKan AI identificere hadeful tale på produktionsskala?
SessionXII (12 hearing)
Convened3 jul. 2026
Previously ruledIN_RESEARCH (May '26) → YES (May '26) → YES (May '26) → YES (May '26) → YES (May '26) → YES (May '26) → YES (Jun '26) → YES (Jun '26) → YES (Jun '26) → ALMOST (Jun '26) → YES (Jun '26) → YES (Jul '26)
Presiding JudgeHon. E. Dijkstra-Patel
II. Cumulative Tally Across Sessions

Across 12 sessions, 31 jurors have heard this case. Combined tally: 27 YES · 3 ALMOST · 1 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 2 — 0 — 0, the panel returns a verdict of JA, with verdict confidence of 93%. The court so orders.

IV. Udtalelser fra dommerpanelet
Nævning I JA

"Leading models (e.g., Perspective API, proprietary systems) detect hate speech at production scale with measured accuracy >90% on standardized benchmarks like HateCheck."

Nævning II JA

"AI models can classify text as hate speech"

Individuelle nævningers udtalelser vises på originalengelsk for at bevare bevismæssig præcision.

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

Hvad publikum mener

Nej 8% · Ja 79% · Måske 14% 132 votes
Ja · 79%
Måske · 14%
15 days of activity

Diskussion

no comments

Kommentarer og billeder gennemgår admin-godkendelse før de vises offentligt.

12 jury checks · seneste for 1 dag siden
03 Jul 2026 2 jurors · kan, kan kan
27 Jun 2026 1 juror · kan kan
22 Jun 2026 2 jurors · kan, uafklaret uafklaret
16 Jun 2026 2 jurors · kan, kan kan
11 Jun 2026 3 jurors · kan, kan, kan kan
05 Jun 2026 4 jurors · kan, kan, uafklaret, kan uafklaret
31 May 2026 3 jurors · kan, kan, kan kan
26 May 2026 3 jurors · kan, kan, kan kan
20 May 2026 3 jurors · kan, kan, kan kan
15 May 2026 3 jurors · kan, uafklaret, kan uafklaret
12 May 2026 3 jurors · kan, kan, kan kan status ændret
11 May 2026 2 jurors · kan, kan ikke uafklaret status ændret

Hver række er et separat jurytjek. Nævninger er AI-modeller (identiteter holdt neutrale med vilje). Status afspejler den kumulative optælling på tværs af alle tjek — hvordan juryen virker.

Flere i Ethical

Har du en vi gik glip af?

Tilføj et udsagn til atlasset. Vi gennemgår ugentligt.