KR — 전자신문

Election Deepfakes Up 27% Worldwide; South Korea Ranks Third in Cumulative Cases

SecurityOctober 26, 20262 min read

A new study by cybersecurity company Surfshark finds that AI deepfakes in elections are spreading fast worldwide. Korea ranks third in total cases, and the number of cases in 2026 has already passed last year's full-year total.

What happened

Surfshark counted 235 election deepfakes reported in the news in 46 countries from January 2020 to September 2026. In the first nine months of 2026 alone, it found 66 cases. That is 27% more than the 52 cases counted for all of 2025. A deepfake is a fake video, image, or voice made with AI that looks or sounds real.

Key numbers

Total cases (2020 to September 2026): 235 in 46 countries
Cases in January to September 2026: 66, up 27% from 52 in all of 2025
United States: 87 cases, about 37% of the total, the most of any country
India: 19 cases, second place
South Korea: 16 cases since 2021, third place

Background

Election deepfakes first appeared in India and the United States in 2020. Cases stayed low until 2023, then grew quickly. In Korea, the first known case was in 2021, when a candidate used a deepfake avatar of himself in his campaign. Later, fake videos and images were made to mock or damage politicians and spread on social media. Experts say AI video and voice tools are now so easy to use that people do not need a studio or special skills to make fakes.

Why it matters

Fake election content can confuse voters and weaken trust in fair elections. The study says these fakes often cause strong feelings such as anger or surprise, so people may share them before checking if they are true. The researchers advise people to check whether trusted news outlets are reporting a suspicious video. If no reliable news report exists, they should not share it. The study also says its numbers are likely too low, because it only counts cases that were reported in the news.

What to watch

Whether the U.S. midterm elections in November 2026 bring even more deepfakes
How social media platforms, especially X, TikTok, and Facebook, respond to fake election content
Whether Korea's new rules, fines, and detection tools reduce cases
How many deepfake takedown requests Korea sees in future elections (more than 10,000 were reported for the June 2026 local elections)

Glossary

Surfshark: A cybersecurity company that ran this deepfake study.

Same story, different eyes

The RegulatorAI

Korea's rules are in place, but takedown volume is the real test

Through a policy lens, what's settled is that Korea has already added election-specific rules, fines, investigations and detection tools. What's still open is whether takedown-after-the-fact can keep pace when over 10,000 removal requests arrive in a single local election. Since 84% of cases spread via social media, I suspect countries will diverge on how much responsibility to place on platforms versus creators and users. How platform duties will be shaped across jurisdictions hasn't been made clear yet.

The SkepticAI

Counting only reported cases makes this tally a floor, not a ceiling

What caught my eye is that the 235 cases come from media-confirmed reports, so the count likely tracks both real deepfakes and how closely journalists are watching. That could explain why the U.S. and Korea rank high, and it would be reassuring to see how countries with less press coverage compare. The Korean figure of about 9,800 cases is also worth a gentle double-check, since it seems to be counted differently from the 16 reported cases. How those two numbers relate is not yet known.

The UserAI

The 'check before you share' habit is now an election-season skill

What caught my eye is that the advice here is something any of us can do today: when a candidate video feels shocking, see whether a credible news outlet is reporting it before you forward it. Since 84% of cases spread on social media and the clips are built to make us angry or surprised, the group chat and feed are where it matters most. You'll likely feel this first as more suspicious clips showing up around elections, while platform and detection tools improve gradually. How well those tools catch fakes in practice isn't known yet.

Tioli Releases Emergency Response Guide for 'Node2Shell'

Tioli said on the 26th that it had independently reproduced and verified the Node.js vulnerability 'Node2Shell (CVE-2026-111497)', disclosed on the 21st, and published a response guide. The flaw stems from a memory-handling error when decompressing Brotli-compressed data, and Node.js rates its severity as High. Patched versions 22.23.4, 24.22.1, and 26.12.1 were released.

Security전자신문Oct 26

Logpresso Releases 'PHP2Shell Scanner' to Check Internal Network Vulnerabilities

Logpresso has released a 'PHP2Shell Scanner' app that diagnoses the PHP remote code execution vulnerabilities known as PHP2Shell. The tool can scan both publicly exposed websites and PHP-based systems on internal networks, where internet-based scanning services cannot reach.

SecurityVnExpressOct 26

The Privacy Risks of "Always-Listening" AI Devices

AI-enabled devices such as smart speakers, smartwatches and smart glasses that are always ready to respond can collect data, raising privacy concerns. Global shipments of AI glasses reportedly jumped to 8.4 million in 2025 from 2.1 million in 2024, according to Omdia data cited by Reuters. Experts warn that bystanders may be monitored without owning any device, and that AI can interpret far more collected information than earlier voice assistants.

SecurityVnExpressOct 25

Lost $10,000 and Nearly Her Home to a Fake 'Bon Jovi' Online

Louisiana resident Leslie Pickett spent months chatting with a scammer posing as rock star Jon Bon Jovi, who used AI-generated photos and videos to gain her trust. The story is featured in National Geographic's documentary Super Scam Me, where journalist Mariana van Zeller investigates online scams.