
South Korea's National Court Administration reported 239,145 hacking attempts targeting the court network from January to August 2024, a 2.4-fold increase from 101,205 cases in the same period last year, with March recording the highest monthly total at 48,263 attempts.
AI-generated summary
The National Court Administration provided data to Rep. Kim Dong-ha of the Democratic Party showing hacking attempts on South Korea's court network from January to August 2024, comparing figures to previous years.
SEOUL, Oct. 4 (Yonhap) -- About 240,000 hacking attempts targeting the court network were reported between January and August this year, about 2.4 times the figure from the same period last year, court data showed Sunday.
According to the National Court Administration data submitted to Rep. Kim Dong-ha of the ruling Democratic Party, hacking attempts totaled 239,145 cases over this period, compared with 101,205 cases during the same period last year and 141,528 cases for all of last year.
By month, March saw the highest number of such attempts with 48,263 cases, followed by May with 39,915 cases and June with 34,083 cases. The annual totals in recent years were 142,514 in 2021, 126,212 in 2022, 114,070 in 2023 and 98,171 cases in 2024, according to the data.

Hackers using AI have attacked from all directions, from commercial banks to savings banks, capital companies, and mutual finance companies, leaking a large amount of customer information, and the financial authorities are expected to acknowledge the limitations of existing security capabilities and push for a complete overhaul of the financial sector's security system.

According to the Epoch AI report, AI inference costs are falling by about 47% per quarter, or one-third of the cost per year, which is four times faster than DNA analysis. The cost of OpenAI's 'o3' model was 30 cents per doctoral level exam question, but after 18 months, the cost of 'GPT-5.6 Luna' was reduced to 0.04 cents, 1 in 725.

As personal information leaks continue to occur in communication, distribution, and online platforms, as well as OTT, payment, beauty platforms, and major commercial banks, sensitive life information such as name and contact information, as well as income, loan limit, and consultation and treatment information, is being exposed. In particular, recent banking incidents have raised the possibility of attacks using artificial intelligence, raising concerns that if leaked information is combined, it could lead to secondary damage such as customized phishing.

David Robinson, a former Open AI safety employee, criticized AI companies for focusing on speed rather than safety, and that Open AI's method of improving when it discovers a problem increases the risk of failure. He warned that as AI gets smarter, it could behave differently during testing and deployment, arguing that it needs layers of safety measures like a nuclear power plant.

As AI competition moves beyond answers to the AI โโagent stage that performs actual work, domestic companies such as KT, Raon Secure, and LG U+ are rapidly expanding the scope of AI use to the fields of work automation, security, lightweighting, and scientific research.

From January to August of this year, there were 239,145 hacking attempts targeting the court computer network, which is 1.7 times the total number of incidents last year, with the highest monthly number of 48,263 in March. The National Court Administration submitted the data to the National Assembly, and as a result of its own audit of personal information leaks due to past hacking incidents, it was found that measures such as severe admonitions were taken.