Artificial intelligence in breast oncology

Abstract Artificial intelligence (AI) is based on complex artificial neural networks, characterized by layered network architecture, parallel processing of large data sets and iterative algorithms for processing large data sets. AI-assisted screening studies have demonstrated non-inferior diagnostic performance, reduced human workload by up to 70%, and reduced recall rates by 25% compared to human double reading. Natural language models promise high accuracy in advising on breast cancer prevention (80%), guiding tumor boards for personalized treatment decisions (50–70%), and planning oncoplastic or radiotherapy treatment for standard cases (72%), but AI sometimes produces errors and fails in complex cases. The main technical advantage of AI is that it can perform routine tasks faster and with fewer errors than humans. This is relevant for scheduling, summarizing reports, recording services for billing and quality assurance. The main concerns in healthcare are the quality of training data, the stability of AI systems, cybersecurity, liability and transparency. Currently, human experts still outperform AI in most areas. AI self-correcting algorithms and the alignment of AI constructed goals with human ethics are imperative to prevent patient harm. The ability of AI to uncover hidden patterns in multi-omics, immune regulation and tumor defense, as well as to develop new drugs, will advance the integrative fight against breast cancer.