A poem
Blooms ━
Universe blooms in every newborn twilight.
Petals rain in veils across the ether.
Cosmos laces outward, unveiling dreams.
Its nascent forces resonate across space, through realms unseen, calling to us, revealing everything.
Have you seen this passage before? Could it be the very first thing you saw on the opening screen? Perhaps most people found its unexpected content puzzling, or simply moved on to the next screen. This passage is an acrostic poem spelling out ‘UBIENT PRIVATE CLOUD INFRASTRUCTURE’ not ‘UPCI’, while also expressing Ubient’s ideals. There is nothing difficult about it. It is simply meant to give your mind a moment to breathe. Now, having taken that moment, shall we give some thought to something important?
How do you protect your sensitive information?
Even in the Age of AI, Personal Privacy Still Matters
Even in the age of AI, personal privacy remains extremely important. One might even say it has become more important. Cloud frontier AI models sometimes offer people ease and convenience, but they also often collect a great deal of information, and may use users’ data for service improvement and model training. Some AI companies even directly ask users to be careful, noting that some conversations may be seen by reviewers and telling them not to enter confidential information. [Source 5]
The Gap Between the Convenience of AI and Awareness of Personal Data Security
As AI advances, we have entered an era in which people have begun to rely on AI for a wide range of their own tasks, including tasks that used to take a great deal of time. But while enjoying this convenience and the benefits of advanced technology, how far has people’s awareness of protecting their privacy and personal information actually kept pace?
How Much Are Real Personal Data Breaches Increasing?
Rather than data measuring users’ security awareness itself, the sections below review the backdrop to that question: the current state of personal data breaches and infringements, using official statistics and overseas reports.
Case 1
Real data shows that the risk to personal information is not merely hypothetical. According to “2025 Personal Data Breach Report Trends and Investigation/Disposition Cases” (Original text: 「2025년 개인정보 유출 신고 동향 및 조사·처분 사례」), released on May 15, 2026 by the Personal Information Protection Commission (PIPC) and the Korea Internet & Security Agency (KISA), personal data breach reports in 2025 totaled 447, up 45.6% from 307 in 2024. Among the causes of breaches, hacking accounted for the largest share at 62% (276 cases), a sharp rise from 171 cases the year before. The PIPC cited malware such as ransomware and web shells, exploitation of web vulnerabilities such as SQL injection, and abnormal access to administrator pages as the main hacking patterns.
Source: Personal Information Protection Commission and Korea Internet & Security Agency, “2025 Personal Data Breach Report Trends and Investigation/Disposition Cases,” May 15, 2026. [Source 1]
Case 2
What is more noteworthy is that this upward trend did not end with 2025. According to a press release the PIPC published on October 2, 2026, breach reports received in the first half of 2026 alone totaled 432, already close to the full-year 2025 total of 447. The PIPC explained that as AI technology spreads, attacks targeting personal information are becoming more sophisticated and intelligent, and personal data breaches and infringements are increasing rapidly. It also stated that incidents continue to occur at businesses holding large volumes of personal data, such as telecommunications, platform, and retail companies, affecting millions to tens of millions of data subjects.
Source: Personal Information Protection Commission, press release “Investigating Personal Data Breaches and Infringements Faster, Fairer, and More Transparently” (on the “Plan for Rapid Investigation of Personal Data Breach and Infringement Cases and Strengthening Investigative Capacity”), October 2, 2026. [Source 2]
Case 3
This problem is not limited to Korea’s domestic breach-report statistics. Verizon’s “2026 Data Breach Investigations Report,” published in 2026, covers events from late 2024 to late 2025 and analyzed more than 31,000 real-world security incidents, of which more than 22,000 were confirmed data breaches, involving organizations in 145 countries. Verizon stated that this is the largest number of breaches it has examined in a single report, but because this reflects the size of the analyzed sample, it is difficult to read it as a measure of the global volume of breaches. Among breaches with an identified entry point, exploitation of software vulnerabilities accounted for 31%, overtaking credential abuse (13%) as the most common initial access vector, and ransomware was involved in 48% of all breaches. In addition, the median threat actor researched or used AI assistance in 15 different documented attack techniques, and most AI-assisted development of malware and tooling was associated with already well-known attack techniques.
Source: Verizon, “2026 Data Breach Investigations Report” and Executive Summary, 2026. [Source 3]
Case 4
The European Union Agency for Cybersecurity (ENISA)’s “ENISA Threat Landscape 2026” analyzed 8,257 incidents recorded between January and December 2025 affecting EU Member States and EU-based organizations. DDoS attacks, which are low in impact, made up 51.3% of recorded incidents, but ENISA assessed that the threat causing the greatest damage in the short term is financially motivated activity, led by ransomware. Financially motivated activity accounted for about 29% of all recorded incidents, and among the recorded claims within it, ransomware deployments made up 47.3% and data breaches 36%. However, the data breach figure is based on sale claims posted on cybercriminal forums, and ENISA stated that it could not verify their authenticity. In addition, in 2025 more than 48,000 new vulnerabilities (CVEs) were published, a 22% increase over the previous year. ENISA observed that attackers mainly use AI as a tool to augment their existing capabilities, while also analyzing that the adoption of AI systems in enterprise environments is creating a new attack surface, and that attacks targeting digital supply chains and third-party dependencies continue.
Source: European Union Agency for Cybersecurity (ENISA), “ENISA Threat Landscape 2026,” September 22, 2026. [Source 4]
Case 5
These sources differ in scope (Korea, worldwide, the EU), period, and methodology, so they cannot be added together into a single figure or compared directly. Even so, both Korea’s official report statistics and the overseas reports show that breaches targeting digital assets, including personal information, are a real and significant threat. Their diagnoses of AI differ in emphasis. The PIPC pointed to the growing sophistication of attacks as AI spreads; Verizon noted that AI is increasing the speed and scale of already known attack techniques; and ENISA noted that AI serves as an aid for attackers while the AI systems being adopted can themselves become a new attack surface. In particular, in the PIPC’s domestic statistics, personal data breach reports were 167 in 2022, 318 in 2023, 307 in 2024, and 447 in 2025. They dipped slightly in 2024 and then rose sharply in 2025, and 432 were received in the first half of 2026 alone. This makes it hard to view the protection of personal data as solely the responsibility of service providers.
According to recently published data, personal data breach reports, as shown above, have on the whole risen sharply over the past several years. It is nothing short of a serious problem. If such news keeps repeating, more people may come to feel fatigued about security. The best form of security can be said to be protecting oneself.
A Local Network You Can Control
Ubient researches easy and systematic management tools for the privacy and security issues that have become more important in modern times, together with personal network users as well as various local network users. Infrastructure, as it is generally called, has mostly been introduced and maintained by companies. Or, only those with extensive IT knowledge or strong interest manage to set up even a small local infrastructure.
A Personal Privacy Infrastructure Anyone Can Build
However, ordinary users inevitably have relatively little such knowledge and interest. Ubient aims to find ways for such users, too, to easily build a personal privacy network infrastructure, as long as they take an interest and make up their minds.
A Local Network Brand for Individual Users, Not a Company
Ubient is not a company. Ubient is a local network infrastructure brand for individual users. Unlike projects in which many people gather to form a team and carry out research and development, progress will inevitably be slower. Even so, Ubient plans to carry out many ideas and projects in order to work together with people who share various commonalities and have been left in the blind spots of their time and society.
Primary Sources
- [Source 1] Personal Information Protection Commission and Korea Internet & Security Agency, “2025 Personal Data Breach Report Trends and Investigation/Disposition Cases” (원문: 「2025년 개인정보 유출 신고 동향 및 조사·처분 사례」), May 15, 2026 (press release: 「개인정보위, 2025년 개인정보 유출 신고 및 조사·처분 사례 분석결과 발표」)
- [Source 2] Personal Information Protection Commission, press release “Investigating Personal Data Breaches and Infringements Faster, Fairer, and More Transparently” (원문: 「개인정보 유출·침해사고, 더 빠르게 더 공정하게 더 투명하게 조사」), October 2, 2026
- [Source 3] Verizon, “2026 Data Breach Investigations Report” and Executive Summary, 2026
- [Source 4] ENISA, “ENISA Threat Landscape 2026,” September 22, 2026 (analysis period: January 1 to December 31, 2025)
- [Source 5] Google, Gemini Apps Privacy Hub
