In 2022, breast cancer was the most frequently diagnosed cancer in women globally, with an estimated 2.3 million new cases and 666,000 deaths. Better tools to find and treat it matter, and artificial intelligence (AI) is one of the fastest-moving. What it can do today, what is not yet proven and what this means for patients.
What is "AI" in this context?
In breast care, AI is mostly software that can learn patterns from very large collections of medical data such as mammograms, tissue slides and genetic profiles. It has been trained on thousands of examples and can highlight suspicious areas, rank cases by risk, or even predict how a tumor might act.
Finding cancer earlier
AI has made the greatest strides in screening. Radiologists reading mammograms sometimes miss cancers or raise false alarms. In many programs two radiologists read every scan, which takes up time.
The first results from actual screening programs in Europe are encouraging:
A large Swedish trial found that AI-assisted reading reduced the screen-reading workload of radiologists by around 44%. A German program saw cancer detection increase by about 18% when AI supported the reading. Other studies found AI-assisted reading was noninferior or modestly better than standard double reading in detection. AI is also being tested in 3D mammography and MRI.
Understanding the tumor
If cancer is found, doctors will want to know the grade and subtype so they can figure out the best treatment. AI tools can look at digital images of tissue slides and estimate these features, and even predict the risk of recurrence. Some research tools are designed to flag patients who may be able to avoid chemotherapy, and others to identify those at higher risk. Most of this is still in the research phase.
Personalizing treatment
Researchers are developing models that can forecast whether a tumor will respond to treatment before surgery, possibly sparing some patients from ineffective and toxic therapy. Early work also tests combinations of drugs on a patient's own tumor sample, and one platform adjusts doses for each individual patient. They are mostly in trials, not in routine care.
AIs are being trialed as chatbots for genetic counseling. A chatbot and a human genetic counselor were similar in satisfaction and knowledge in a small randomized trial of newly diagnosed patients. That's promising for increasing access, but it was a small study.
Support after treatment
Recovery is not just the cancer itself; anxiety, body image, fatigue and quality of life all matter. There are a few projects trying out apps and platforms that use AI to spot emerging issues and recommend support. Others are trying out tools that display likely cosmetic outcomes to patients before they pick a surgery. Many of these trials have not reported results and AI is also being used to accelerate surgical planning for breast reconstruction.
The honest limits
Most evidence is retrospective. Many studies test AI on old data. Only a handful of prospective studies in real screening exist, mainly in Europe, and we know little about whether AI improves long-term survival. Bias. If an AI system is trained mostly on one population, it may perform worse for others.
The "black box" problem. Many systems can't explain their reasoning, and many are proprietary, which makes independent checking harder. Cost and access. AI may save money over time, but it requires infrastructure and training. Hospitals in low-resource settings may be left behind. Regulation. Several mammography AI tools are cleared in the US as decision-support aids, but approval pathways are still catching up with the technology.
Will AI replace doctors?
Not in the foreseeable future. The more realistic picture is that AI handles repetitive tasks like triage and measurement, while clinicians make decisions, interpret results in context, and talk with patients.
What this means for you
Routine screening should be maintained. The current evidence for AI is promising; however, the standard remains regular mammography performed by trained radiologists.
A clinician still reviews your results if your screening program employs AI, as it is a support tool.
Inquire with your care team about the use of AI in your care and the individuals who are accountable for making decisions.
This post is for education and is not medical advice. Talk to your healthcare provider about your own situation.
References:
Chua, B.N., Thng, D.K.H., Ho, D. et al. Artificial intelligence for breast cancer management. Communications Medicine 6, 79 (2026). https://doi.org/10.1038/s43856-025-01342-3
Global burden of female breast cancer: new estimates in 2022, temporal trend and future projections up to 2050 based on the latest release from GLOBOCAN. Journal of the National Cancer Center 5(3), 287–296 (2025).
Lång, K. et al. Artificial intelligence-supported screen reading versus standard double reading in the Mammography Screening with Artificial Intelligence trial (MASAI): a clinical safety analysis. The Lancet Oncology (2023).
Eisemann, N., Bunk, S., Mukama, T. et al. Nationwide real-world implementation of AI for cancer detection in population-based mammography screening. Nature Medicine (2025). https://doi.org/10.1038/s41591-024-03408-
