Hello! I'm Shen, Data Scientist and AI Safety Researcher. I design large-scale randomized controlled trials (RCTs) and other human uplift studies to evaluate the biological capabilities of frontier AI. With an interdisciplinary background in clinical trials, computer science, applied mathematics and human subject research, I help ground AI capability evaluations in realistic experiments under field conditions.
At Active Site, I work across the entire research pipeline: from IRB review, pre-registration, trial operations, blinded statistical analysis plan (SAP), to final results. I care very strongly about rigorous evidentiary practices in human uplift studies. I design our RCTs using the same standards as clinical trials (SPIRIT, CONSORT, ICH). Blinding, randomization, the use of ICH E9 (R1) addendum estimands, we do it all!
In our most recent work, we conducted a large-scale (n=153) RCT to evaluate whether AI improves novice performance on a set of laboratory tasks that collectively model an influenza viral reverse-genetics workflow. That work is now available as "Measuring Mid-2025 LLM-Assistance on Novice Performance in Biology" (arXiv:2602.16703), and was cited by the New York Times, the Economist, and in the model cards of frontier AI labs.
I was previously a researcher at the Johns Hopkins Bloomberg School of Public Health, where I worked on SKOAP – one of the largest clinical trials in osteoarthritis in North America. It was at Hopkins where I learned how to use REDCap, the importance of outcome adjudication, registered SAPs, and the best practices of a gold-standard, confirmatory trial.
On the technical side, I was a member of Columbia University's PRAISE lab, an AI research group dedicated to the interpretability of ML models in medicine. Led by Dr. Ansaf Salleb-Aouissi, I worked on using influence functions to improve the interpretability of SVM models in predicting pre-eclampsia. I have some experience with PyTorch, and familiarity with the Python ML ecosystem.
I have a BA in Philosophy, Literature, and the Natural Sciences from St. John's College, a BSc. in Computer Science (1st Class Honors) from the University of London, and I studied mathematics at Columbia University. I am currently pursuing a (part-time) MSc. in Applied and Computational Mathematics at Johns Hopkins University, with a focus on causal inference and the design of experiments.
I am one of the authors of the aforementioned novice uplift randomized controlled trial (Hong et al. 2025), and I want to clarify that while the study did not detect a statistically significant difference on the full model reverse-genetics workflow, nonetheless we did find uplift on component tasks such as mammalian cell culture!
Here, a reasonable takeaway is that tacit knowledge in biology is not a strong barrier for uplift — instead, participants in the AI arm performed much better with AI assistance, even on tasks that we would consider to be tacit-knowledge heavy.
An important caveat is that the study was conducted mid-2025, using the then-frontier models of that generation but more importantly, with a participant pool that is representative of the average level of AI proficiency of that period. As time goes on, secular trends in AI skill development will likely make people better at using models.
Finally — I'm curious to hear what's the second RCT that you mentioned in your post? I'd be grateful to receive the link to it!
Thank you for sharing this remarkably well-written overview of magnanimity (megalopsychia). If I may add to your reflections on If you take your morality to extremes you may not be prepared for what you find there, I think you should read Søren Kierkegaard's work Fear and Trembling. In it, he explores what it means to pursue righteousness and moral glory to it's extreme through the story of Abraham and Isaac. It's been a while since I had re-read his work, and while I fear to paraphrase him poorly, but essentially Kierkegaard makes a distinction between the magnanimous man, and the "merely" tragic hero.
Kierkegaard explores magnanimity through the story of the Binding of Issac, and how Abraham's sacrifice of his son is a paradigmatic example of magnanimity and it's moral terror. For those unfamiliar with the story, Abraham was asked by God to sacrifice his son, Isaac, and Abraham complies. Kierkegaard opens the monograph by asking "What sort of man could sacrifice his only son in such a manner?" i.e. pursue righteousness and moral virtue to the most terrific of extremes? He reminds us that Abraham's decision to sacrifice his son was not one that happened 'in the spur of the moment,' but it must have came after several days of thought and preparation — especially since the journey to the mountain was one that must have taken several days:
Kierkegaard makes a case that the magnanimous person is somehow different from the tragic hero. The tragic hero is one who "merely" pursues ordinary virtues to their extremity. In Kierkegaard's words:
This could describe the someone who donates 10% of their income for the certain good, they whom we may applaud and appreciate unreservedly, even if we may not have the strength to pursue their virtues to the same extent. However, a magnanimous man is somehow by nature different — they are, properly speaking, an object of horror.
And is it not the case that your example of Beatrice and Cornelius Boeke one that invokes a similar sense of horror? Perhaps that is the difference between mere moral excellence, and true magnanimity. It is like the difference between a person who donates 10 percent of their income, versus someone who takes a vow of poverty — gives up their worldly possessions, a person who donates their kidney, who sacrifices the very possibility of an ordinary life for that higher thing.
I think you're right to intuit that at a first glance, a life of magnanimity can be a very unhappy existence. Although, I would also bring up Socrates' argument from The Gorgias, where he claims that it is better the suffer injustice than to act injustly, where a just (i.e. magnanimous life) is happy for it's own sake.
I'm grateful for your posting of this essay on righteousness and megalopsychia. While I also hesitate to opine on it, as I do not believe I have came upon any conclusions myself, it is clear to me that these matters are important for us to contemplate if we seek to pursue a moral life, irrespective of our underlying values. You've given me much to think about!
Thank you Toby! St. John's was a wonderful experience. It was by far the best way one could study philosophy in the modern age. For those that are unfamiliar; St. John's is a 4-year liberal arts college with a unique, 'Great Books' pedagogy. Essentially, students learn how to read and translate Ancient Greek and French. Then, over the course of our four years, we read the entire Western canon of philosophical works, in their unabridged form. Students discuss their readings through weekly seminars, with their peers and professors. Here's the reading list, for those that are interested.
The St. John's program is probably the only philosophy program where one still engages in the practice of the activity of philosophy, i.e. contemplation — as opposed to the mere scholarship and historiography of philosophy, i.e. "what did Aristotle write?"
My time there was able to substantively influence my own moral values and ethical understanding, and as I'm pursuing a more technical vocation (currently doing my MSc. in Applied Maths and working in biosecurity research!), I'm grateful to have had this background to help me navigate the process.
So anyways! Thank you for the kind welcome, Toby. I've been spending the past few days reading voraciously on the forum, and I hope to author my own post soon!
I had the pleasure of meeting members of the SecureBio team in Cambridge, MA during their August Summer Party. I can attest that they are all fantastic individuals, and are driven and passionate about their work. I think they're making a great impact!
Hello everyone! My name is Shen, and I'm a philosophy and computer science graduate. I completed my BA in Philosophy and Classics at St. John's College, Annapolis MD, and my BSc. in Computer Science at the University of London. Now I'm pursuing a MSc. in Applied and Computational Mathematics. I currently work as a researcher at the Johns Hopkins School of Public Health.
I'm interested in Effective Altruism, because of the community's contributions to public health, particularly in the form of effective health interventions. I'm always excited about ways to help improve global health outcomes, and I'm keen to learn more about the EA community and it's philosophy of "doing good better."
I'm not quite formally acquainted with the EA philosophy, but I hope to learn more about it with time. I've already enjoyed reading through the EA Forum and it's topics on RCTs and public health/global development interventions, and I hope to learn more as time goes on!