AI Library
The Age of Autonomous Scientific Discovery
Kim Kyung-jin, Attorney at Law
AI Scientists and Self-Driving Labs
This book follows how AI scientists and self-driving labs are changing the way science generates and verifies claims. It covers literature-based discovery, natural-language protocols translated into robot commands, multi-agent research systems, closed-loop laboratories, materials search, the verification gap, chains of evidence, research harnesses, journal ethics, and legal responsibility.
AI Library
A New Era of Life Sciences Opened by Artificial Intelligence
Structural Proteomics, Genomic Foundation Models, Autonomous Laboratories, and Global Governance
Kim Kyung-jin, Attorney at Law
This book is a research volume compiled with artificial intelligence. A human selected the materials and structured the work, while AI models drafted the sentences and cross-checked the facts.
AI Library
The Double Structure of Digital Sovereignty
Europe’s Departure from Palantir and the Chains of American Big Tech
Kim Kyung-jin, Attorney at Law
This is a record of 2026, when European intelligence agencies and defense ministries began removing analytics tools from America’s Palantir. It covers the replacement decisions made by France’s General Directorate for Internal Security (DGSI), Germany’s Federal Office for the Protection of the Constitution (BfV), and the Netherlands Ministry of Defense; the incident in which US export controls severed an ally’s ac…
New English Edition
Artificial Intelligence in Horticulture
Kim Kyung-jin, Attorney at Law
Across five chapters and ten sections, this book examines computer vision for crop diagnosis, harvesting robots and autonomous field systems, smart greenhouses and digital twins, precision irrigation and supply-chain quality control, high-throughput phenotyping, and predictive breeding.
New English Edition
Artificial Intelligence in Food Crop Agriculture
Kim Kyung-jin, Attorney at Law
Across six chapters and eighteen sections, the book examines digital agricultural infrastructure, remote sensing, crop diagnosis, yield forecasting, precision irrigation, genomics, molecular breeding, agricultural robotics, climate-smart agriculture, and global food security.
New English Edition
The Future of Forestry and Agroforestry
Kim Kyung-jin, Attorney at Law
Driven by Artificial Intelligence and Digital Innovation
Across five chapters and fifteen sections, the book follows satellites, drones, LiDAR, digital twins, forest-specific language models, wildfire and pest forecasting, forestry robotics, agroforestry, timber traceability, and forest carbon markets.
New English Edition
Smart Livestock Farming: AI Enters the Barn
Kim Kyung-jin, Attorney at Law
Sensors listen, cameras watch, and artificial intelligence helps farmers decide.
Across five chapters and fifteen sections, the book follows precision livestock farming from animal health and reproduction to robotic milking, virtual fencing, digital twins, methane reduction, welfare, and data ownership.
Table of Contents
Han Dong-hoon, Busan Buk-gu Gap: A Record of the 100 Days Before and After the Election (Mar. 26-Jul. 3, 2026)
Kim Kyung-jin
Table of Contents and 13 sections
From March 26 to July 3, 2026, this record follows the spring after expulsion, the Busan Buk-gu Gap by-election, victory as an independent, and the first bill submitted in the National Assembly.
Table of Contents
Artificial Intelligence and Medicine
Kim Kyung-jin, Attorney at Law
AI in clinical care, hospitals, education, and research
AI in medical imaging, risk prediction, treatment planning, hospital operations, education, and research, with patient safety, privacy, and accountability.
[AI Library] 21 NHS Data Issues and AI Ethics
Demis Hassabis, Father of Google's Artificial Intelligence
Part 7. AI for Science
21 NHS Data Issues and AI Ethics
Kim Kyung-ran, Kim Kyung-jin
In 2016, the data agreement with NHS Royal Free Hospital raised concerns over access to 1.6 million medical records without patient consent and the privacy implications that followed. In the cold London air of 2016, Demis Hassabis and his longtime friend and DeepMind co-founder Mustafa Suleyman were facing an unexpectedly brutal ordeal. Before the cheers over artificial intelligence's victory on the Go board had even died down, a complex and sensitive real-world problem had caught them by the ankles. It was not a technical challenge.
It was a matter of 'trust' and 'ethics,' the most fundamental values of human society that no code could solve. The story began with the launch of DeepMind Health, DeepMind's ambitious medical project. Hassabis and Suleyman had identified healthcare as the field where artificial intelligence should bring change first.
Their logic was clear. If DeepMind's technology could predict protein structures and solve complex games, it should also be used to save patients' lives. Their first partner was Royal Free Hospital, located in north London.
But the goal hit trouble from the start. New Scientist, a prominent British science magazine, published an investigative report raising questions about the data-sharing agreement between DeepMind and Royal Free Hospital. The contents were shocking.
DeepMind had gained access to 1.6 million patient medical records from Royal Free Hospital to develop an app. The data included not only records of currently admitted patients but also records spanning the previous five years. The bigger issue was the scope of this data.
It covered patients' names, dates of birth, and addresses, along with blood test results, pathology records, and radiology images. When it became known that the data could even include extremely private information such as HIV status, drug overdose history, and abortion records, British society erupted. Patients
had no idea that their most intimate information had ended up in the hands of a subsidiary of Google, one of the largest tech companies in the world. Protests poured in: 'We never consented to this.' For Hassabis, it was a painful blunder.
He had always insisted that 'artificial intelligence must be used for the benefit of humanity,' yet he could not escape the criticism that he had overlooked the rights of the very individuals who make up that humanity. DeepMind pleaded its case. They argued the data was used solely to develop and test an app designed to assist patient care, not for research or commercial purposes.
They also explained that the data was encrypted and managed separately, never combined with Google's other data. The public remained cold. It was a fear born of information asymmetry.
People felt a vague yet concrete dread that DeepMind's algorithms might learn from their medical records and later use that knowledge to raise insurance premiums or hurt their job prospects. The UK Information Commissioner's Office (ICO) launched an investigation, and in July 2017, it ruled that Royal Free Hospital had violated the UK Data Protection Act by failing to adequately inform patients about the data sharing. For Hassabis, this case meant more than a legal problem.
It was a clear example showing that the Silicon Valley ethos, 'Move fast and break things,' could not apply in the medical field, where human life and dignity are at stake. Good technical intentions were not enough. DeepMind learned at great expense that innovation without procedural legitimacy and transparency would inevitably meet social resistance.
Hassabis fell into deep reflection. As an engineer, he had pursued efficiency, but as a leader, he had to shoulder social responsibility. The 1.6 million records were not just numbers.
Each one represented a human life, a person who had suffered, healed, or sometimes faced death. DeepMind later reorganized its ethics board and fundamentally reviewed its approach to data access. This would have a decisive influence on how carefully they handled data when undertaking large-scale scientific projects like AlphaFold in the years that followed.
The Streams App from DeepMind Health and the Detection of Acute Kidney Injury. The Streams app, which stood at the center of the controversy, was ironically a highly practical and intuitive piece of software that used almost no artificial intelligence technology. But the app's origins and functions precisely illustrated the intersection between the reality facing the British healthcare system (NHS) at the time and DeepMind's pursuit of 'solving real-world problems.' The story begins with the frustrating reality that nephrologists at Royal Free Hospital were dealing with every day. Acute Kidney Injury (AKI) is a common complication among hospitalized patients, a condition in which kidney function deteriorates suddenly. AKI can be easily treated with fluids or medication adjustments if caught early, but late detection can lead to permanent kidney damage or death.
In the UK alone, statistics showed that more than 40,000 patients per year died preventable deaths due to AKI. The problem was the speed of information. In 2016, many NHS hospitals still relied on paper charts, pagers, and fax machines.
When a patient's blood test results came back, the lab might spot warning signs, but it could take hours or even a full day for that information to reach the attending physician. Doctors had to manually sift through chart after chart for each patient, and in the meantime, kidneys were silently failing. Mustafa Suleyman wanted to fix this inefficiency with technology.
The principle behind the Streams app was straightforward. It connected to the hospital's systems and, when signs of AKI appeared in a patient's blood test results, such as a sharp rise in creatinine levels, it immediately sent an alert to the attending medical staff's smartphones. Just as stock traders receive real-time notifications about sudden market movements, doctors could now be instantly aware of a patient's crisis.
Streams organized a patient's medical history, current medications, and vital signs into an easy-to-read interface. Doctors no longer needed to dig through piles of paper charts. They could assess a patient's condition at a glance on their smartphone screen and issue orders. The DeepMind team shadowed doctors and nurses like ghosts to develop this app, thoroughly analyzing their movements and workflows. It was the moment when user experience design met the urgency of the medical frontline.
Technically, Streams was not cutting-edge AI like deep learning or large language models. It was closer to a data visualization tool that adapted existing medical algorithms for mobile use. Hassabis and Suleyman made an important decision here.
'Rather than showing off flashy AI technology, what matters is building a tool that saves lives on the ground right now.' That was their pragmatic approach. They decided that before applying AI, they needed to open up the data pipeline first. The results from Streams were immediate.
During the pilot at Royal Free Hospital, the time needed to identify AKI patients dropped from hours to seconds. Medical staff saved an average of two hours per day on administrative tasks, freeing that time for patient care. Faster emergency response showed the potential to improve patient survival rates. Nurses and doctors on the ground welcomed the app, calling it 'a revolution in the healthcare system.'
But these shining results were overshadowed by the data privacy controversy described earlier. DeepMind argued that it needed data from all patients, not just those with AKI, to verify that the app worked properly. The public and regulators saw it differently: they had invaded personal privacy by burning down the barn to catch a flea. In the end, the Streams project left a painful lesson that no matter how useful a technology is or how good its intentions, it cannot be sustained if it steps outside social consensus and legal boundaries.
In 2019, when the DeepMind Health team was absorbed into Google Health, the operational control of the Streams app passed to Google. This sparked another round of controversy. The early promise that 'DeepMind's medical data will be permanently separated from Google's advertising business' appeared to have been broken.
Streams proved how powerful information technology could be in a medical setting, while simultaneously becoming a landmark symbolizing the distrust and conflict that big tech companies face when entering the public healthcare domain. For Hassabis, Streams was both a success and a wound, an inoculation that posed questions he would have to answer before the coming age of AI. Achievements and Limits of Medical AI: The Eye Disease Diagnosis AI, the Nature Medicine Paper, and the Collaboration with Moorfields Eye Hospital. In the summer of 2016, Demis Hassabis knocked on the door of Moorfields Eye Hospital, a world-renowned ophthalmology center located on London's Old Street.
He had just gone through a harsh initiation with the data controversy at Royal Free Hospital, but he did not stop. This time was different. He approached more cautiously, more precisely, and above all, more transparently.
His partner was Pearse Keane, a young and passionate ophthalmologist at Moorfields. Dr. Keane was fighting a desperate situation. As society aged, the number of patients with conditions like age-related macular degeneration (AMD) and diabetic retinopathy was surging.
The hospitals had Optical Coherence Tomography (OCT) machines to diagnose these conditions, but there were far too few specialist eyes to read the thousands of three-dimensional scan images pouring out. Patients had to wait weeks or months for a diagnosis, and many lost their sight in the interim. Dr. Keane told Hassabis:
'We are drowning in a flood of data. Can your AI save us?' Hassabis saw this proposal as the perfect opportunity to fulfill DeepMind's mission: 'Solve intelligence to advance the world.'
But this time, he took every precaution with data handling. DeepMind received 14,884 anonymized OCT scan datasets from approximately 15,000 patients at Moorfields Hospital. Drawing on the lessons of the Royal Free incident, they clearly defined the purpose and scope of the data and conducted thorough de-identification.
The DeepMind research team designed a two-stage AI model based on this data. The first neural network analyzed OCT images and performed precise segmentation of eye tissue and lesions. Like drawing a map, it separated the layers of the retina and abnormal areas pixel by pixel.
The second neural network used this map to classify disease types and assess urgency. In August 2018, the joint research by DeepMind and the Moorfields team was published in Nature Medicine, one of the world's leading medical journals. The AI model matched or outperformed Moorfields' top specialists in diagnosing more than 50 types of eye disease.
The accuracy exceeded 94 percent, and the diagnostic error rate was 5.5 percent, lower than the 6.7 percent average error rate among eight human experts. The system excelled at triage, identifying urgent patients and prioritizing their care.
The real reason this study earned praise from the academic world was not just the high accuracy. Hassabis and the research team worked to address the chronic 'black box' problem of medical AI. Conventional deep learning models would produce results but could not explain 'why' they made a particular judgment, which bred distrust among doctors.
DeepMind's model was different. Along with the diagnosis, the AI marked lesion areas on the OCT scan image in distinct colors. Doctors could see exactly which parts the AI had examined to reach its conclusion, allowing them to verify and trust the AI's judgment.
This was a model case of 'Explainable AI (XAI)' applied in a clinical setting. DeepMind's system was not meant to replace doctors but to hand them a powerful 'digital microscope.' Doctors could use the AI's initial findings to make faster and more accurate final diagnoses. This project became a textbook example of how technology companies and medical institutions should collaborate.
Through this work, Hassabis demonstrated to the world that DeepMind was not just a company that played Go well but one capable of solving real challenges in science and medicine. The Moorfields project was a signal flare, restoring the trust DeepMind had lost and reigniting hope that AI could contribute to human health. Lessons on the Ethical Use of Medical Data. The successful collaboration with Moorfields Eye Hospital and the painful mistake at Royal Free Hospital left Demis Hassabis, DeepMind, and AI researchers worldwide with profound lessons about medical data ethics.
The realization was this: 'Data is not just a resource. It is a token of trust.' The first lesson concerns the value of transparency and consent. In the Royal Free case, DeepMind claimed it had obtained approval from the NHS's information governance officer, known as the Caldicott Guardian. But blanket data access without explicit consent from individual patients failed to win social acceptance.
This coincided with the implementation of the European Union's General Data Protection Regulation (GDPR) in 2018, which fundamentally shifted awareness of data sovereignty. From that point on, DeepMind invested heavily in creating patient participation groups for all its medical projects and explaining in plain language how data would be used. No matter how
excellent a technology may be, if the process behind it is opaque, society will reject it. They felt that truth in their bones. The second lesson is the clarity of purpose. When developing the Streams app, DeepMind collected comprehensive medical records for the specific purpose of detecting acute kidney injury.
This drew criticism for violating the Data Minimization Principle, which holds that only the minimum data necessary to achieve the stated purpose should be used. By contrast, the Moorfields project limited its scope to eye scan images and directly related diagnostic data, all thoroughly anonymized. This suggested that data collection for AI training should not follow a 'the more, the better' approach but a 'precision targeting' one.
The third lesson is the tension between public interest and private corporate interest. The NHS is a public healthcare system funded by British taxpayers. Its data is a public asset belonging to the people.
Is it legitimate for a subsidiary of Google, the world's largest for-profit company, to use that data for free to develop algorithms and then sell the resulting technology back to the NHS or to other countries for profit? This question remains open. Hassabis had to consider open-sourcing DeepMind Health's technology or finding models that returned benefits to the public. It meant that AI companies must become parties to a social contract, not just technology suppliers.
The final lesson is 'human-centered technology.' In both Streams and the eye diagnosis AI, the ultimate decision-maker was the doctor. AI had to stay in the role of a capable investigator who gathers evidence, not a judge who renders verdicts.
The explainability shown in the Moorfields project was an essential condition for machines to collaborate with humans without sidelining them. DeepMind proved that clinical utility and explainability matter just as much as technical accuracy. For Hassabis, medical AI was a game in an entirely different dimension from the chessboard or the Go board.
A Go board is a closed system with 19 lines, clear rules, and only victory or defeat. But a hospital is an open system tangled with incomplete information, patient suffering, privacy, and legal regulations. Through countless trials and errors in this open world, DeepMind learned that 'solving intelligence' does not simply mean improving the performance of algorithms
but also includes helping that intelligence land safely within society. These lessons would become DeepMind's ethical compass when they later tackled the biological challenge of protein structure prediction with AlphaFold, and when they discussed the prospect of more powerful artificial general intelligence (AGI). 'The most powerful technology demands the highest level of responsibility.'
That was the discovery Hassabis made in the wards of Royal Free and the consulting rooms of Moorfields, a finding as valuable as any Nobel Prize. NHS Royal Free London Hospital
Kim Kyung-jin
Attorney · Former Member of the National Assembly · AI Policy Researcher
© 2026 Kim Kyung-jin. All rights reserved.