Realtime AI News
South Korea's financial sector hit by fresh cyberattacks as AI pentest tool draws scrutiny
Daishin Securities, Mirae Asset Global Investments and crypto exchange Bithumb reported cybersecurity incidents this week, with Daishin confirming that data on 38,490 customers was exposed. The wave comes a month after earlier attacks on South Korean financial firms that researchers linked to a suspected China-based hacker using an AI penetration testing tool called Artex.

South Korea's financial sector has been hit by another wave of cyberattacks, TechCrunch reported on October 28, with brokerages Daishin Securities and Mirae Asset Global Investments and crypto exchange Bithumb among those affected. The incidents come roughly a month after a series of attacks on South Korean financial institutions, widening concern beyond the country's banking systems.
Daishin has confirmed a breach that exposed customer data. According to local news cited by TechCrunch, the attack exposed the personal information of 38,490 customers, including names, email addresses and mobile phone numbers. The company has notified affected customers, is cooperating with authorities and will compensate for losses as required by law.
Mirae and Bithumb reported separate cybersecurity incidents on Wednesday, though neither has confirmed that customer data was taken. TechCrunch said South Korea's Financial Services Commission, Daishin Securities, Bithumb and Mirae Asset Global Investments did not immediately respond to requests for comment.
The more telling context is the tooling. Security researchers linked the September wave of attacks on South Korea to a suspected China-based hacker who relied on an AI penetration testing tool called Artex. Officials are still investigating the root cause of the latest breaches, and investigators have not established a connection between the incidents.
Regulators are already responding. South Korea's financial regulator called for stronger cybersecurity measures following the string of attacks on financial firms and government agencies, and the country plans security checks on dozens of private-sector IT systems. A day earlier, Japan reported an incident involving GMO Internet, after a series of attacks on its own financial institutions.
Read together, the episode shows AI acting as an amplifier of attack surface: penetration testing tools are built for defenders to probe their own systems, but when misused they lower the barrier to finding exploitable holes. The targets also shifted from banks toward brokerages and crypto exchanges, where third-party integrations, customer data pools and trading systems make for concentrated value.
What to watch next is the investigative record: whether Artex is confirmed in this round, what common weaknesses the Financial Services Commission's checks turn up, and whether Korean and Japanese regulators build a joint response to a cross-border pattern. TechCrunch notes that a link between the incidents remains unproven.
Why it matters
The incidents turn AI-assisted hacking from an abstract warning into measured data loss, and give South Korea's regulator a concrete case for auditing private-sector IT systems, especially third-party integrations and customer data pools.
Nearby Updates
All10/29, 00:05
GenLayer Labs warns x402 agent payments have no way to dispute a charge
GenLayer Labs is warning that payments made by AI agents over the x402 protocol have no built-in dispute, chargeback or refund path once a transaction settles. As machine-to-machine purchasing grows, the missing recourse layer is becoming one of agentic commerce's most practical gaps.
10/28, 23:00
NVIDIA's RTX Spark laptops gain new models, local AI models and agent GPU controls
NVIDIA says RTX Spark Windows laptops are now shipping from ASUS, Dell, HP, Lenovo, Microsoft and MSI, starting at $2,599 with up to 128GB of unified memory. Alongside the hardware, NVIDIA added the open-weight Nemotron 3.5 Transcribe speech model, GPU resource controls for background agents in the NVIDIA App, and DLSS 4.5 support for Phantom Blade Zero.
10/28, 22:58
NVIDIA open-sources Nemotron 3.5 Transcribe with a promptable speech stack
NVIDIA published Nemotron 3.5 Transcribe on October 28, an open-weight model that folds transcription, speaker attribution, translation and audio question answering into one checkpoint covering 40 languages. The release ships with the Granary v2 data pipeline and fine-tuning recipes, and NVIDIA reports first place among open-weight models on the Hugging Face English ASR leaderboard at 5.33% word error rate.
10/28, 22:57
Diffusion Intelligence's Clover-35B takes diffusion models into agent work
Diffusion Intelligence released Clover-35B-A3B-Preview on October 28, a 35-billion-parameter diffusion language model with roughly 3 billion active parameters aimed at coding and agent tasks. On the 49-task WindTunnel web-agent benchmark its median execution time was 1.52 seconds against 3.21 seconds for a Jev plus Mercury 2.5 combination, with a 98.6% success rate.