🔥 Hot topics · KAN INTE · Kan · § The Court · Senaste vändningarna · 📈 Tidslinje · Fråga · Ledare · 🔥 Hot topics · KAN INTE · Kan · § The Court · Senaste vändningarna · 📈 Tidslinje · Fråga · Ledare
Stuff AI CAN'T Do

Kan AI generera en anpassad sociala medier-djupfalsk video av en specifik person som säger vad som helst ?

Vad tycker du?

Spridningen av deepfake-teknik har demokratiserat desinformation och möjliggjort hyperrealistiska videoförfalskningar. AI-system kan nu skapa skräddarsydda falska innehåll som är anpassade efter en individs röst, sätt att uttrycka sig och sammanhang. Detta undergräver förtroendet för digital media och möjliggör trakasserier, utpressning och politisk manipulation. Plattformar har svårt att upptäcka och mildra sådana hot i stor skala.

Background

The proliferation of deepfake technology has democratized misinformation, enabling hyper-realistic video forgeries. AI systems can now create bespoke fake content tailored to an individual’s voice, mannerisms, and context. This undermines trust in digital media and enables harassment, blackmail, and political manipulation. Platforms struggle to detect and mitigate such threats at scale.

Current systems can generate highly realistic “talking head” videos that sync a person’s face to a new voice and script. Producing a custom deepfake that convincingly depicts a specific individual saying anything requires both a clear, high-quality image or short video of the target and a robust audio sample that captures their vocal patterns. Techniques like diffusion models (e.g., Stable Diffusion Video, Runway Gen-2) and GAN-based methods (e.g., StyleGAN, DeepFaceLab) have advanced to the point where short clips with lip-sync and facial movements are possible; yet artifacts, lighting mismatches, and temporal inconsistencies still reveal synthetic origins to trained observers. Ethical and legal frameworks, including detection tools and content provenance standards such as C2PA, are being developed but do not yet prevent misuse entirely. Generative AI in this domain continues to evolve rapidly, posing ongoing challenges for verification and trust.

— Enriched May 12, 2026 · Source: U.S. Department of Commerce

Status senast kontrollerad June 27, 2026.

📰

Galleri

In the Court of AI Capability
Summary of Findings
Verdict over time
May 2026May 2026May 2026May 2026May 2026Jun 2026Jun 2026Jun 2026Jun 2026Jun 2026
Sitting at the Bench Filed · jun 27, 2026
— The Question Before the Court —

Kan AI generera en anpassad sociala medier-djupfalsk video av en specifik person som säger vad som helst?

★ The Court Finds ★
Reaffirmed
Ja

Juryn fann ett tydligt jakande svar.

Ruling of the Bench

The jury found unopposed evidence that today’s AI can stitch together a believable deepfake video of a named individual speaking new words, pulling from hours of public footage without further human help. With no dissent to temper their confidence, the panel returned a unanimous verdict that the capability exists now. The bench declares: “The court sees a face, the court sees a voice—verdict for yes.”

— Hon. E. Dijkstra-Patel, Presiding
Jury Tally
2Ja
0Nästan
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 · 83%
Session III · May 2026 Ja · 85%
Session IV · May 2026 Ja · 84%
Session V · May 2026 Ja · 79%
Session VI · Jun 2026 Ja · 77%
Session VII · Jun 2026 Ja · 82%
Session VIII · Jun 2026 Nästan · 90%
Session IX · Jun 2026 Ja · 95%
Case № 7829 · Session X
In the Court of AI Capability

The Case File

Docket № 7829 · Session X · Vol. X
I. Particulars of the Case
Question put to the courtKan AI generera en anpassad sociala medier-djupfalsk video av en specifik person som säger vad som helst?
SessionX (10 hearing)
Convened27 jun 2026
Previously ruledIN_RESEARCH (May '26) → YES (May '26) → YES (May '26) → YES (May '26) → YES (May '26) → YES (Jun '26) → YES (Jun '26) → ALMOST (Jun '26) → YES (Jun '26) → YES (Jun '26)
Presiding JudgeHon. E. Dijkstra-Patel
II. Cumulative Tally Across Sessions

Across 10 sessions, 27 jurors have heard this case. Combined tally: 24 YES · 2 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. Uttalanden från rätten
Jurymedlem I JA

"Generating realistic deepfake videos of specific persons is achievable with models like DeepFaceLab, FaceSwap, or diffusion-based video editing."

Jurymedlem II JA

"Advanced deep learning models can generate realistic videos"

Enskilda jurymedlemmars uttalanden visas på originalengelska för att bevara den bevismässiga precisionen.

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

Vad publiken tycker

Nej 35% · Ja 57% · Kanske 9% 23 votes
Nej · 35%
Ja · 57%
60 days of activity

Diskussion

no comments

Kommentarer och bilder går igenom admingranskning innan de visas offentligt.

10 jury checks · senaste för 20 timmar sedan
27 Jun 2026 2 jurors · kan, kan kan
22 Jun 2026 1 juror · kan kan
17 Jun 2026 2 jurors · oavgjort, kan oavgjort
11 Jun 2026 3 jurors · kan, kan, kan kan
06 Jun 2026 2 jurors · kan, kan kan
31 May 2026 2 jurors · kan, kan kan
26 May 2026 4 jurors · kan, kan, kan, kan kan
20 May 2026 4 jurors · kan, kan, kan, kan kan status ändrad
15 May 2026 4 jurors · kan, oavgjort, kan, kan oavgjort
12 May 2026 3 jurors · kan, kan inte, kan oavgjort

Varje rad är en separat jurykontroll. Jurymedlemmar är AI-modeller (identiteter avsiktligt neutrala). Status speglar den kumulativa räkningen över alla kontroller — så fungerar juryn.

Fler i technology

Har du en vi missat?

Lägg till ett påstående i atlasen. Vi granskar veckovis.