KR — ZDNet Korea

AI Era Privacy Future: Calls to Unify "Collection and Provision" Under "Processing"

SecurityNovember 1, 20262 min read

Experts in Korea met to discuss how to update privacy law for the AI era. Many said the current system, which relies on user consent, no longer works well and should be reorganized around the idea of "processing" personal data.

What happened

At a public forum held on August 1 at Yonsei University in Seoul, panelists and audience members proposed changes to Korea's personal data rules. The main idea was to merge the separate rules for collecting and sharing data into one framework called "processing." They also suggested wider use of "legitimate interest," a legal basis that lets a company use personal data without asking for consent, as long as the user keeps the right to say no.

What is "legitimate interest"?

Legitimate interest is a legal reason that lets a company handle personal data without consent. It applies only when the data is needed for the company's goal, the person's rights do not clearly outweigh the company's interest, and the use stays within a reasonable limit. Right now, Korean law reads this idea strictly. One panelist said that allowing it for the benefit of a third company would be hard under current law.

Key numbers

Only 51% of people said they read all the consent details before deciding.
88% said they had agreed to consent forms without really reading them.
The European Union gave Korea an "adequacy decision" in December 2021, which allows data to move safely between Korea and the EU.
For important terms, the law requires text at least 9 points in size, and it may also require different colors and boldness.

Why it matters

Consent forms are the main way companies get permission to use personal data today. If most people click "agree" without reading, the system may not protect anyone well. AI adds new risks, because data used to train a model can stay inside the model even after it is deleted from a database. Experts warned that there is no perfect solution, so the government must set careful rules and quick ways to help people who are harmed.

What to watch

Whether lawmakers merge the collection and sharing rules into one "processing" framework.
How the idea of shared processing and wider legitimate interest is written into law.
Whether industry groups create standard consent forms for each business sector.
How the government handles data that AI models keep after it is deleted from a database.

FBI: North Korea behind multibillion-dollar crypto theft from Binance

According to the FBI, North Korean hackers from the group TraderTraitor carried out the multibillion-dollar attack on the cryptocurrency exchange Binance, which the FBI put at more than $6 billion. The FBI asked companies to block deposits and withdrawals from certain addresses to hinder laundering of the stolen assets. Binance says it has covered 90 percent of the stolen sum and that first withdrawals resumed on Sunday morning.

Security전자신문Oct 31

AI Security Race Shifts from 'Detection' to 'Automated Response'… Market Growth Also Accelerates

AI-driven cyberattacks are shifting the competitive focus of the global security market from threat detection and analysis to real-time blocking and automated response. Fortinet, Tenable, and F5, which recently reported earnings, said their security approaches have changed as AI attacks spread, and they expect security investment to grow as a result. F5 said its AI-based web application firewall blocked multiple attack variants in its own tests without new blocking rules after a recent PHP vulnerability disclosure.

Securityheise onlineOct 31

Popcorn Time: Abandoned Update Servers Threatened Mac Users

Security firm Eye Research reports that attackers could have taken over the abandoned update servers of Time4Popcorn, an alternative Popcorn Time client, and run arbitrary code, even with root rights on macOS, on affected Windows, macOS and Android systems. Eye Research instead took over the servers itself and pushed an update that stopped the insecure updater from running. The firm estimates that 250,000 to 350,000 macOS machines were affected, and the details were presented at the hack.lu conference.

SecurityZDNet KoreaOct 31

Column on Security: The Fall of Human-Centered Cybersecurity

This ZDNet Korea column argues that this autumn's wave of breaches across Korea's financial sector, which the authorities say has hit 16 firms, exposed the limits of human-operated, patch-centered defense. It contrasts these attacks with historical examples such as the Opium War and calls for zero trust, AI-assisted defense, and leadership that measures security by how quickly intrusions are detected and stopped.