Greek-Australian teen builds AI model

Michaela Loukas, an 18-year-old Greek-Australian student in Sydney, has drawn attention after The National Herald reported on June 1 that she developed an artificial-intelligence model designed to detect breast cancer. The report said Loukas is now studying a double honors degree in Engineering (Bioinformatics) and Advanced Science, after building the project as part of her Higher School Certificate work. The same project won her the 2025 Young Scientist of the Year title from the Science Teachers’ Association of New South Wales, according to the association and other Australian Greek-community outlets. Social media posts amplified the story this week, but the reporting available publicly describes a student research project rather than a peer-reviewed medical tool. ### Who is Michaela Loukas, and what was she recognized for? The National Herald identified Loukas as a Greek-Australian teenager and Sydney university student who was “recently recognized” for developing an AI model that can detect breast cancer. The outlet said she had previously taken part in coding programs and competitions and was named Young Scientist of the Year in 2025. (thenationalherald.com) The Science Teachers’ Association of New South Wales said on November 28, 2025, that Loukas, from Marist Catholic College Penshurst in southern Sydney, won its top student award for a project titled “Assessing the Accuracy and Interpretability of a Recurrent Neural Network for Breast Cancer Classification and Molecular Subtyping using Ribonucleic Acid Sequencing Data.” The association said the annual competition drew more than 500 entries. (thenationalherald.com) ### What does the model actually analyze? Loukas told The National Herald that the model uses “genetic information from tissue biopsies” to detect and better understand breast cancer. She said the system was trained to recognize gene-expression patterns that differ between healthy breast cells and cancerous ones, and across subtypes of breast cancer. (stansw.asn.au) The National Herald also quoted Loukas saying she focused on interpretability so the model could identify the specific genes it relied on. “This kind of interpretability is essential if A.I. is to be safely used in real-world healthcare,” she said. That description places the project in the category of computational analysis of biopsy-derived RNA data, rather than a screening device already used in clinics. (thenationalherald.com) ### What evidence is public about the model’s performance? A December 2025 report published by NewsMoCo, which summarized the same student project, said Loukas’s model achieved 98% accuracy in distinguishing cancerous from healthy breast tissue and 70.6% accuracy in identifying four common breast-cancer subtypes. The report also said Loukas described the work as not yet ready for clinical application. (thenationalherald.com) That same report said the project highlighted the gene transferrin as potentially significant in breast-cancer development. Reuters could not independently verify the methodology, dataset, or performance claims from a peer-reviewed paper in the material reviewed for this article, and no journal publication was identified in the available sources. (newsmoco.com) ### Is this the same as a validated breast-cancer screening product? Publicly available reporting does not show that Loukas’s model has been validated in peer-reviewed clinical research or approved for medical use. The National Herald described the work as a student project and quoted Loukas discussing how AI could be used to “catch cancer earlier” and classify it more precisely. (newsmoco.com) BreastCancer.org said in a May 2026 explainer that AI shows potential to help radiologists detect cancerous tissue more quickly and accurately, but it described the field broadly rather than Loukas’s project specifically. The University of Melbourne said in March 2026 that its BRAIx tool was developed on mammograms from nearly 400,000 women and tested on almost 96,000 more, illustrating the scale of validation typically described for medical AI systems moving toward real-world use. (thenationalherald.com) ### What happens next for Loukas’s work? The Science Teachers’ Association of New South Wales said selected finalists from the 2025 Young Scientist Awards will go on to represent the state at the National iCubed Science Awards. The available reports do not specify whether Loukas has published the breast-cancer project in a scientific journal or entered clinical testing with a hospital or company partner. (breastcancer.org) The National Herald said Loukas is now pursuing bioinformatics and advanced science studies in Sydney. Any next formal milestone for the project would most likely be a publication, competition result, or institutional partnership, but no such step was named in the sources reviewed as of June 2, 2026. (thenationalherald.com) (stansw.asn.au)