Researchers at the University of Technology Sydney (UTS) have developed an artificial intelligence model that analyses DNA to identify genetic factors associated with ischaemic stroke, potentially improving personalised prevention and treatment.

Stroke is the world’s second-leading cause of death. Ischaemic stroke, the most common type, occurs when a blood clot blocks blood flow to the brain.

The research team identified 18 genetic variants associated with an increased predisposition to ischaemic stroke. The findings could help doctors assess inherited risks more accurately and provide families with better information about their chances of developing the condition.

“When someone has a stroke, particularly at a young age, they often want to know whether other family members are also at risk,” said Kairui Guo, the study’s corresponding author. “These new AI tools will provide the ability to better help families understand that risk.”

Previous research has suggested that genetic factors account for approximately 38% of an individual’s stroke risk. Guo said genetics had traditionally received less attention in stroke risk assessments, despite their potential significance.

The researchers developed a model based on transformer technology, the same general approach used in large language models such as ChatGPT. It was trained on genetic variations to identify patterns and relationships linked to ischaemic stroke.

The model highlighted genes including EDEM2, which is active in the brain cortex, and KANK2, which is expressed in whole blood, as important contributors to genetic predisposition.

Researchers evaluated the system using two major medical datasets containing more than 40,000 people combined. It successfully divided individuals into low, medium and high genetic risk categories.

Over time, people classified in the high-risk group were 21% more likely to experience a first ischaemic stroke than those placed in the low-risk group.

The findings could help address limitations in existing methods for estimating inherited disease risk. Scientists have commonly used polygenic risk scores, which calculate genetic susceptibility by combining the effects of numerous genetic variations.

However, conventional scores can overlook interactions between genetic variations and the significance of their positions within a person’s DNA sequence. The new AI approach aims to account for such relationships when assessing risk.

Jie Lu, director of the Australian Artificial Intelligence Institute, said AI could identify associations between genes and diseases and support personalised risk assessments to help prevent serious medical events.

The next step is to develop a multimodal AI system that combines genetic information with other sources of data, including medical records, lifestyle surveys, scans and medical imaging.

Lu said genetic information was only one part of the overall picture. Combining it with established risk factors such as blood pressure, age, atrial fibrillation, type 2 diabetes and cholesterol levels could improve monitoring and prevention.

The researchers’ findings point towards a potential role for AI in more personalised stroke care, although further work would be needed before the technology could be adopted in routine clinical practice.

The European Trade Union Confederation (ETUC) is calling for menopause to be formally recognised in European Union workplace legislation, arguing that inadequate support is costing women earnings, career progression and pension income.

The confederation, which represents 45 million members across 41 European countries, wants employers to provide better workplace adjustments and protections for women experiencing menopause symptoms.

“Women’s work-related health risks remain largely ignored by the law, which remains orientated around male-dominated jobs,” said ETUC Deputy General Secretary Isabelle Schömann.

“It’s high time that the law caught up with the reality of the whole workforce today and the risks facing them,” she added.

The campaign comes as the European Parliament examines a report urging the EU to treat menopause as an economic, employment and social issue. The European Commission has acknowledged menopause in its wider gender equality agenda but has stopped short of proposing dedicated legislation or a separate EU strategy.

Menopause typically occurs between the ages of 45 and 55 and can affect women’s ability to work. Around 85% of women experience at least one symptom, while approximately 30% report symptoms that affect their capacity to work. These can include fatigue, sleep disruption, hot flushes, joint pain, anxiety and difficulty concentrating.

Research cited by the ETUC found that women’s earnings in Norway fell by 7.4% by the fourth year after a menopause diagnosis. Reduced working hours, sick leave and missed promotions contributed to the decline.

“Loss of pay, lost promotions, and lower pensions is the price being paid by millions of women for the failure to adapt work to the menopause,” Schömann said.

EU institutions face pressure to act

The European Commission’s Gender Equality Strategy 2026-2030, published in March, includes commitments to exchange national best practices on menstrual and menopause poverty. It also plans a study for 2028 examining the economic and social benefits of closing the women’s health gap, including menopause.

However, in November 2025, Equality Commissioner Hadja Lahbib said the Commission was not planning a dedicated menopause strategy, pointing instead to research supported through Horizon 2020 and Horizon Europe.

Spanish Socialist MEP Lina Gálvez is leading a draft European Parliament report titled Menopause beyond a health issue: turning menopause into an economic, working and social priority.

The report calls for a dedicated EU strategy, expanded research, public campaigns to challenge stigma, stronger workplace support and better access to healthcare. It is expected to receive further discussion in Parliament next year and has not yet become adopted EU policy.

Unions seek practical workplace changes

The ETUC wants workplace risk assessments to cover factors such as temperature, ventilation and facilities. It is also calling for flexible working hours, remote working options, management training and physical adjustments, including temperature controls and ergonomic equipment.

The confederation wants menopause policies to be negotiated with trade unions, implemented and monitored. It is also seeking free menstrual and menopause products in workplaces.

Schömann said the way work is organised can significantly affect women’s experience of menopause, adding that unions are already negotiating collective agreements covering workplace adjustments and flexible arrangements.

The campaign highlights growing pressure on EU institutions and employers to address menopause as a workplace issue, rather than treating it solely as a private medical concern.

Artificial intelligence is helping health authorities respond more quickly to the Ebola outbreak in the Democratic Republic of Congo (DRC), cutting the time needed to prepare a situation report from a full working day to about an hour. However, experts warn that concerns over transparency, reliability and accountability must be addressed before the technology becomes more widely used in healthcare.

Since May, health workers in the DRC have been responding to an outbreak of the Bundibugyo strain of the Ebola virus. Teams visit communities, provide advice, record deaths and recoveries, and refer people with suspected symptoms to medical facilities.

Information collected by these workers has been digitised and analysed using Anthropic’s Claude artificial intelligence model. The initiative involved the Coalition for Epidemic Preparedness Innovations (CEPI), the World Health Organization’s Regional Office for Africa and the country’s National Institute of Biomedical Research (INRB).

The system helps identify people who may have been exposed to the virus, trace contacts and locate potential outbreak hotspots. This information allows authorities to prepare medical supplies and position resources in areas considered at higher risk.

Marie Roseline Darnycka Belizaire, director for emergency preparedness and response at the WHO Regional Office for Africa, said the technology had reduced the time required to produce situation reports to around one hour.

Speaking at the World Health Summit, she said AI could help health authorities assess the spread of an outbreak more quickly and direct medical assistance where it was needed most.

Supporting overstretched healthcare systems

Thomas Zapata, director of the WHO Academy in Lyon, said AI could help address growing demand for healthcare amid persistent shortages of medical staff and resources.

However, he cautioned against treating the technology as a solution in itself, arguing that it must be integrated into existing health systems and support the work of healthcare professionals.

AI could reduce administrative workloads, improve diagnostic processes and assist clinical decision-making, allowing medical staff to spend more time examining patients, listening to their concerns and arranging appropriate treatment.

Cecilia Mbaka, a senior technical adviser for health and emergencies at aid organisation CARE, said the technology could be particularly useful for repetitive tasks and information processing. She stressed that its value should be measured by whether it improves care and gives medical professionals more time to attend to patients.

In the DRC, however, shortages of equipment and limited public confidence in healthcare facilities remain major challenges. Epidemiologist Maximilian Gertler has estimated that around half of deaths in villages occur outside medical facilities, highlighting difficulties in accessing formal care.

Germany recently increased its assistance to the country by €21 million.

Questions over transparency and reliability

Despite the potential benefits, experts warn that many AI systems cannot clearly explain how they reach particular results.

Thomas Thesen, a professor of medical education and informatics at Dartmouth University’s Geisel School of Medicine, said AI models were highly effective at identifying patterns but often struggled to explain what those patterns meant.

This creates concerns when AI systems offer diagnoses or treatment recommendations. Medical professionals need to assess whether a recommendation is appropriate for an individual patient and understand its limitations. Questions also remain over who should be held responsible when an AI-generated recommendation is wrong.

AI should not replace medical education

Experts also warn that excessive reliance on AI could weaken the training of future healthcare professionals.

Thesen cited research suggesting that students who used AI chatbots to find information developed more superficial knowledge than those who relied on conventional search methods.

He said independent learning and mental effort were essential to retaining information, particularly in medicine, where professionals must make decisions even when technological assistance is unavailable.

The Ebola response demonstrates how AI can help process information quickly during emergencies. Experts say its wider use in healthcare will depend on maintaining human oversight, protecting patients and ensuring that technology strengthens, rather than replaces, professional judgement.