News
AI used by English councils downplays women’s health issues – study
Artificial intelligence used by more than half of England’s councils is downplaying women’s physical and mental health issues, raising gender bias concerns in care decisions.
Research found that when using Google’s AI tool “Gemma” to generate and summarise the same case notes, words such as “disabled”, “unable” and “complex” appeared significantly more often in descriptions of men than women.
The London School of Economics and Political Science (LSE) study also found that similar care needs in women were more likely to be omitted or described in less serious terms.
Dr Sam Rickman, lead author of the report and a researcher in LSE’s Care Policy and Evaluation Centre, said AI could result in “unequal care provision for women”.
He said: “We know these models are being used very widely and what’s concerning is that we found very meaningful differences between measures of bias in different models,.
“Google’s model, in particular, downplays women’s physical and mental health needs in comparison to men’s.
“And because the amount of care you get is determined on the basis of perceived need, this could result in women receiving less care if biased models are used in practice.
“But we don’t actually know which models are being used at the moment.”
AI tools are increasingly being used by local authorities to ease the workload of overstretched social workers, although there is little information about which specific models are in use, how often they are applied and what impact this has on decision-making.
The LSE research used real case notes from 617 adult social care users, which were inputted into different large language models (LLMs) multiple times, with only the gender swapped.
Researchers then analysed 29,616 pairs of summaries to see how male and female cases were treated differently by the AI models.
In one example, Gemma summarised the male case notes as: “Mr Smith is an 84-year-old man who lives alone and has a complex medical history, no care package and poor mobility.”
The same case notes, with the gender swapped, were summarised as: “Mrs Smith is an 84-year-old living alone. Despite her limitations, she is independent and able to maintain her personal care.”
In another example, Mr Smith was described as “unable to access the community”, while Mrs Smith was said to be “able to manage her daily activities”.
Among the AI models tested, Google’s Gemma created more pronounced gender-based disparities than others.
Meta’s Llama 3 model did not use different language based on gender, the research found.
Rickman said the tools “must not come at the expense of fairness”.
He said: “While my research highlights issues with one model, more are being deployed all the time, making it essential that all AI systems are transparent, rigorously tested for bias and subject to robust legal oversight.”
The paper calls for regulators to “mandate the measurement of bias in LLMs used in long-term care” to ensure “algorithmic fairness”.
There have long been concerns about racial and gender biases in AI tools, as machine learning techniques have been found to absorb prejudices in human language.
One US study analysed 133 AI systems across different industries and found that about 44 per cent showed gender bias and 25 per cent exhibited gender and racial bias.
According to Google, its teams will examine the findings of the report.
The research tested the first generation of Gemma, which is now in its third generation and is expected to perform better, although Google has never stated the model should be used for medical purposes.
An AI tool could help spot breast cancer risk by analysing how individual breast cells behave when squeezed under stress, research suggests.
Researchers at City of Hope and the University of California, Berkeley, created a microfluidic platform that assesses women’s breast cancer risk at the cellular level.
The platform squeezes individual breast epithelial cells, which line breast tissue, to measure how they deform, recover and behave under stress.
Because more than 90 per cent of women do not have a known genetic predisposition to breast cancer or a family history of the disease, the researchers said the approach could help fill a key gap in risk assessment.
Mark LaBarge, professor in the department of population sciences at City of Hope, said: “For women with a known genetic risk factor for breast cancer, there are things you can do like follow a higher-risk screening protocol. For everybody else, you’re left wondering, ‘Am I at high risk?’
“By translating physical changes in cells into quantifiable data, this tool gives women something tangible to discuss with their doctors, not just risk estimates, but evidence drawn directly from their own cells.”
The researchers developed a machine learning algorithm that identifies and measures cells showing signs of accelerated ageing, generating an individual breast cancer risk score.
They said the platform uses simple electronics that could be easy and affordable to replicate on a large scale.
Lydia Sohn, chair in mechanical engineering at UC Berkeley, said: “Our team isn’t the first to measure the mechanical properties of cells; however, other approaches require advanced imaging technology that’s expensive, cumbersome and has limited availability.
“In contrast, MechanoAge uses computer chips that are simpler than an Apple Watch and ‘RadioShack parts’ that are cheap and easy to assemble, potentially making the device highly scalable.”
About 6 per cent of women who develop breast cancer carry known genetic mutations.
For women outside this group, risk is usually estimated indirectly using population models or measures such as breast density, which can both overestimate and underestimate individual risk.
The researchers said there is currently no non-genetic test that can identify women at higher risk of breast cancer.
Screening mammograms can detect cancer only once it has started to grow, but the MechanoAge platform aims to assess risk earlier by looking for physical changes in individual cells.
Using the platform, the researchers found that breast cells appear to have a “mechanical age” separate from a person’s chronological age, based on how the cells respond to stress.
They said this is the first time mechanical age has been quantified in biological cells.
Sohn said: “We learned that the older the mechanical age, as determined by how cells respond to being squeezed through our microfluidic device, the higher the risk for breast cancer.”
In this type of mechano-node-pore sensing, an electrical current is measured across a liquid-filled channel.
As cells pass through, they disrupt the current, generating measurements about their size and shape. By narrowing parts of the channel, researchers squeeze the cells and then measure how long each one takes to return to its normal shape.
The team found that cells from older women were stiffer and took longer to bounce back after being squeezed.
They also identified a subset of younger women whose cells behaved more like those from older women. These cells came from women with genetic mutations linked to a higher breast cancer risk.
The researchers then refined the algorithm to assign a risk score based on the cells’ measured mechanical and physical properties. They said it successfully identified women with known genetic risks.
The team then used it to compare cells from healthy women, women with a family history of breast cancer, and cells taken from the healthy breast of women with breast cancer in the other breast.
LaBarge said: “With accuracy, we were able to figure out which women were at high risk of breast cancer and which women didn’t seem to be.”
The work grew out of more than 12 years of collaboration between the two labs, combining engineering with cancer and ageing biology.
Sohn said: “It’s a true collaboration. We’ve learned a lot from each other.
LaBarge added: “In my view, this is what happens when you have a real collaboration that develops over a long time. This result is not what we imagined at the beginning.”
An experimental drug slowed triple-negative breast cancer in mice by flooding tumour cells with toxic fats.
Triple-negative breast cancer lacks three common drug targets, making it one of the hardest-to-treat and most aggressive forms of the disease.
The compound, known as DH20931, appears to push cancer cells past their limits by triggering a surge in ceramides, fat-like molecules that place the cells under intense stress until they self-destruct.
In lab experiments, the drug also made standard chemotherapy more effective. When combined with doxorubicin, researchers were able to reduce the dose needed to kill cancer cells by about fivefold.
The drug targets an enzyme known as CerS2 to sharply increase production of these lipids and stress cancer cells. Healthy cells, by contrast, showed lower sensitivity to the drug in lab tests.
While the early results are promising, further preclinical and clinical trials would still be needed to determine the safety and effectiveness of DH20931 in humans.
Satya Narayan, a professor in the University of Florida’s College of Medicine, led the study with an international group of collaborators.
The researchers published their results on human-derived tumours on 21 April and presented their findings on combination therapy at the annual meeting of the American Association for Cancer Research in San Diego.
Narayan likened the drug’s effects to a home’s electrical system handling a power surge.
While healthy cells act like a properly grounded and installed circuit, cancer cells are more like a jumble of mismatched wires and faulty fuses. DH20931 overwhelms cells not with electricity, but with fats.
He said: “When that surge goes into the cancer cells, they cannot handle the amount of power they are getting. The fuses burn out, the cell can’t handle the surge and it dies.”
The compound was developed at the University of Florida in the lab of Sukwong Hong.
Hong, now a professor at the Gwangju Institute of Science and Technology in South Korea, created DH20931 as one of many drug candidates tested for efficacy in Narayan’s lab.
In the study, researchers implanted human triple-negative breast cancer tumours into mice and treated them with DH20931.
The drug significantly slowed tumour growth without causing noticeable weight loss or signs of toxicity in the animals. In separate lab experiments, it also showed activity against other breast cancer subtypes.
In addition to increasing lipid levels, DH20931 triggers a second stress signal by flooding cells with calcium.
Together, these effects disrupt the mitochondria, the structures that produce a cell’s energy, ultimately leading to cell death.
Narayan said: “It does not just follow one pathway but it goes through multiple pathways. It’s a two-hit hypothesis.
“These pathways are common in all breast cancer types and other solid tumours, so we think this drug can be useful not only in triple-negative breast cancer but potentially other cancers as well.”
Entrepreneur
Future Fertility raises Series A financing to scale AI tools redefining fertility care worldwide
Future Fertility Inc. has announced the closing of a US$4.1 million Series A financing round.
The round was led by M Ventures (the corporate venture capital arm of Merck KGaA, Darmstadt, Germany) and Whitecap Venture Partners, with participation from new investors Sandpiper Ventures, Gaingels, and Jolt VC.
The financing will accelerate Future Fertility’s commercial expansion into Asia-Pacific and support its entry into the United States, including planned FDA 510(k) clearance for additional products as part of a broader U.S. market entry strategy.
Proceeds will also advance the development of a broader AI platform, from egg assessment through to embryo transfer, designed to support clinicians, embryologists, and patients across the full IVF journey.
M Ventures and Whitecap have supported Future Fertility’s mission to translate AI innovation into meaningful clinical outcomes since the company’s earliest stages.
Oliver Hardick, investment director, M Ventures, said: “Future Fertility is addressing a critical unmet need in reproductive medicine with a differentiated AI platform grounded in clinical data and real-world workflow integration.
“We are excited to continue supporting the company and team because we believe its technology has the potential to improve decision-making for clinicians, bring greater clarity to patients, and help advance a more personalised standard of care in fertility treatment.”
Future Fertility’s AI platform addresses a long-standing gap in fertility care: historically, there has been no objective, clinically validated method for assessing egg quality (Gardner et al., 2025), despite it being one of the most important drivers of reproductive success.
The company’s suite of deep learning tools includes VIOLET™, MAGENTA™, and ROSE™, purpose-built for egg freezing, IVF, and egg donation respectively.
The tools are based on AI models trained and validated on more than 650,000 oocyte images and are deployed in over 300 clinics across 35 countries.
Rhiannon Davies, founding and managing partner, Sandpiper Ventures, said: “The best outcomes in fertility care globally come from better data and smarter tools. Future Fertility understands that, and they’ve built a platform that delivers on it.
“Sandpiper is proud to back a team turning rigorous science into real results for patients and clinicians alike.”
Partnerships with the world’s leading fertility networks – including IVI RMA and Eugin Group across Latin America and Europe, FertGroup Medicina Reproductiva in Brazil, and most recently announced Kato Ladies Clinic in Japan – reflect growing demand for objective, AI-powered oocyte assessment in fertility care. In the United States, ROSE™ is newly available under an FDA 513(g) determination.
Research shows that approximately 50 per cent of IVF patients do not understand their likelihood of success, and many discontinue treatment prematurely, even though cumulative success rates improve significantly with multiple cycles (McMahon et al., 2024).
By delivering earlier clarity on egg quality, Future Fertility’s tools support more informed conversations between clinicians and patients, helping set realistic expectations and guide decisions about next steps.
Future Fertility’s growing evidence base spans seven peer-reviewed publications in Human Reproduction, Reproductive BioMedicine Online, Fertility & Sterility, and Nature’s Scientific Reports, and more than 70 scientific abstracts accepted and presented with partner clinics at conferences worldwide.
Christine Prada, CEO, Future Fertility, said: “Fertility treatment is one of the most emotionally and physically demanding experiences a person can go through.
“Every patient deserves objective data, not just a best guess, to support better decisions at critical moments in their care.
“This funding means we can bring that clarity to more patients, in more countries, at a moment when it matters most.”
Find out more about Future Fertility at futurefertility.com
Entrepreneur3 weeks agoThree sessions that show exactly where women’s health is heading in 2026
Entrepreneur3 days agoFuture Fertility raises Series A financing to scale AI tools redefining fertility care worldwide
Pregnancy4 weeks agoHow NIPT has evolved and what AI NIPT means in 2026
News3 weeks agoTwo weeks left to make your mark in women’s cardiovascular health
Opinion4 weeks agoQ1 momentum: Female founders are advancing, but the system still hasn’t caught up
Menopause2 weeks agoMore research needed to understand link between brain fog and menopause, expert says
Diagnosis2 weeks agoFuture Fertility partners with Japan’s leading IVF provider, Kato Ladies Clinic
News2 weeks agoSelf-employment linked to better cardiovascular health outcomes in Hispanic women