Academic Solidarity e.V.

A study published in Nature Medicine in January 2026 revealed that sleep is not just a moment of rest. Sleep is also a powerful biological signal that can predict future disease risks. Researchers showed that by developing a multi-mode artificial intelligence model called SleepFM, they could predict the risk of more than 130 diseases with high accuracy from sleep recordings of a single night (https://www.nature.com/articles/s41591-025-04133-4).

The study is based on a massive dataset comprising over 585,000 hours of polysomnography (PSG) recordings from approximately 65,000 individuals. PSG is a gold-standard sleep study method that simultaneously records multiple physiological signals, including brainwaves (EEG), heart rate (ECG), respiration, and muscle activity. SleepFM is designed as a foundation model that learns the “language” of sleep by analyzing these different signals together.

One of the most striking aspects of the study is its high predictive accuracy for cardiovascular diseases. The model can predict outcomes such as heart failure, stroke, myocardial infarction, and death due to cardiovascular causes with significant accuracy.

In particular, external validation performed on an independent dataset yielded an AUROC value of 0.88 for cardiovascular death. This value indicates a very strong discriminatory power in clinical predictive models. Similarly, high accuracy values were obtained for stroke and heart failure. It is stated that the combined evaluation of ECG signals and respiratory parameters plays a crucial role in the model’s success.

These findings suggest that physiological signals recorded during sleep can detect cardiovascular risks early on, even before they manifest clinically.

Today, sleep tests in clinical practice are mostly used to diagnose problems such as sleep apnea, insomnia, or excessive daytime sleepiness. However, this study reveals that sleep data has a much broader potential:

These findings mean the following in daily practice:

• A patient’s sleep record for one night can provide predictions about their future cardiovascular disease risk. This can enable early intervention, especially in asymptomatic individuals.
• High-risk patients can be individually referred for more intensive lifestyle interventions, close monitoring, or further investigations.
• In family medicine and cardiology practice, artificial intelligence models integrated into electronic patient records can provide physicians with objective risk scores.
• It can be more rationally determined who should undergo further investigation or which patients should be monitored more frequently.

On the other hand, the study population in this research mainly consists of patients who applied to sleep clinics; therefore, direct generalization to the general population may be limited. In addition, the fact that the decision-making mechanism of the artificial intelligence model is not fully explainable remains a subject of debate regarding clinical acceptance. Therefore, the results should be seen as tools to support physician assessment, not as a replacement for it.

The study is based on retrospectively matching sleep lab data with electronic patient records. However, it is unclear whether all patients who underwent PSG were followed up in the same health system over the long term. The lack of a central national EHR infrastructure in the US means that some diagnoses may have been recorded in different institutions and not reflected in the dataset. This situation may pose a risk of loss to follow-up, particularly in long-term disease predictions, and could impact model performance.

Ultimately, despite the limitations of this research, the researchers emphasize that models like SleepFM can be integrated with sleep data from wearable devices in the future. As smartwatches and home sleep sensors become more prevalent, non-invasive and continuous health monitoring may become a possibility in the near future.

Developed countries are currently facing two major and interconnected demographic transformations simultaneously. On the one hand, the population is rapidly aging, fertility rates are falling, and life expectancy is increasing; on the other hand, these countries are increasingly receiving immigration, and their social structures are becoming more diverse. Although these two processes are often discussed under separate headings, they are in reality parts of a single major transformation that simultaneously challenges the functioning of states, welfare systems, and the social contract.

The issue of aging is reflected most starkly in the numbers. While the proportion of the working-age population is decreasing, the proportion of the retired and elderly population is increasing. This has a direct impact on many areas, including pension funds and healthcare systems. Systems that finance pensions and healthcare expenses with taxes and contributions paid by workers are becoming increasingly fragile as the demographic balance is disrupted. In most European countries and across the OECD, it is predicted that the elderly dependency ratio will increase sharply in the coming decades. Simply put, this means fewer workers will have to finance a greater number of elderly people (https://ec.europa.eu/eurostat/statistics-explained/SEPDF/cache/80393.pdf).

This situation creates significant pressure on pension systems. While measures such as raising the retirement age, increasing contribution rates, or relatively reducing pension payments are technically possible, each generates political and social tensions. Moreover, as the proportion of the elderly population among the electorate increases, implementing such reforms becomes even more difficult. The pension system functions not only as a financial mechanism but also as a relationship of intergenerational solidarity and trust. When this trust is undermined, the social contract itself is called into question.

The impact of an aging population is not limited to pensions. The need for health and long-term care services increases with age, creating permanent burdens on public budgets. At the same time, there is a risk of labor market contraction. When fewer people are working, the potential for economic growth decreases, and financing public services becomes even more difficult. Therefore, many countries are trying to increase women’s employment, encourage older individuals to work longer, and develop a growth model based on productivity increases (https://www.oecd.org/en/publications/2025/11/pensions-at-a-glance-2025_76510fe4/full-report/demographic-old-age-to-working-age-ratio_25476b96.html).

One of the most striking examples of this demographic shift is Japan. Japan has a population structure characterized by low fertility and high life expectancy, resulting in a population that is not only aging but also declining in absolute terms. Projections for 2050 indicate that the country’s social structure, labor market, rural areas, and urban planning will all be reshaped in response to this reality of contraction and aging. The Japanese example also serves as a time machine for Europe; many problems that many countries will face decades from now have already become part of daily life in Japan. In short, Japan is the country that says “The future has arrived” for Europe (https://www.japantimes.co.jp/news/2025/12/29/japan/society/japan-2050-predections-depopulation/).

One of the most important balancing mechanisms developed against aging is migration. Working-age migrants can support the labor market, broaden the tax base, and help address chronic labor shortages in certain sectors. Therefore, Europe, North America, and other developed regions have experienced intense immigration in recent years. In the European Union today, approximately one-tenth of the population consists of people born in non-EU countries. In OECD countries, the number of foreign-born people exceeds hundreds of millions (https://ec.europa.eu/eurostat/statistics-explained/index.php?title=EU_population_diversity_by_citizenship_and_country_of_birth).

However, the issue of migration is often addressed within a narrow and reductionist framework. In public discourse and politics, migration is often discussed through the lens of communities from Africa, the Middle East, and Asia, predominantly identified as Muslim. However, global migration flows are far more complex. Asia, Latin America, and Africa are significant regions of migration, with diverse contexts. Key determinants of migration include numerous factors such as war, political oppression, economic inequality, climate change, education, and job opportunities. Religious identity is only a small and often overemphasized part of this process (https://worldmigrationreport.iom.int/msite/wmr-2024-interactive/).

There is no definitive answer to the question of whether migration is a solution for aging societies. When supported by the right policies, migration can mitigate the economic effects of aging. However, when integration remains weak, migration can generate new social inequalities, exclusion, and political tensions. Without strong policies in areas such as education, language learning, housing, equitable access to the labor market, and combating discrimination, the potential of migration is largely wasted. Therefore, the determining factor is not the quantity of migration, but the extent to which migrants can participate in social and economic life. For countries sending migrants, the picture takes on a different dimension. The forced or semi-forced migration of academics, students, and qualified professionals, in particular, leads to a significant brain drain in these countries. While universities, healthcare systems, and public institutions weaken, receiving countries can benefit from this human resource. However, this process is not inherently fair or efficient. Problems with diploma recognition, precarious employment, and loss of professional status remain common experiences for qualified individuals migrating.

At this point, demographic transformation ceases to be merely a matter of population numbers. It also represents a global redistribution of knowledge production, academic freedom, and institutional capacity. Discussions on aging and migration directly intersect with questions of forced migration, academic solidarity, and equal participation. Receiving societies must learn to live with diversity and support it with a fair integration policy. Societies sending migrants, on the other hand, face the long-term consequences of losing their human resources. How this transformation is managed will determine not only today’s economic balances but also the future of social peace and academic freedom. When managed correctly, demographic transformation can create opportunities for global solidarity; when managed incorrectly, it prepares the ground for long-term crises in both sending and receiving societies. In this context, there can be both a good and a bad scenario:

The good scenario: Developed countries, in response to the aging problem, view the migrants they receive not only as temporary elements filling labor shortages but also as long-term social actors. Effective integration policies, education, and academic development opportunities support the potential of migrants while fostering egalitarian and solidarity-based relations with countries of origin. Diaspora networks, academic collaborations, and return channels contribute to the reconstruction of the human potential lost by migrant-sending societies. This reciprocal interaction forms the basis of a sustainable and stable system for both receiving and sending countries.

Worst-case scenario: When flawed integration policies, economic stress, and identity politics converge, the migrant population becomes the scapegoat for societal problems. Social exclusion deepens, democratic norms weaken, and political polarization increases. Simultaneously, brain drain in sending countries accelerates institutional decay and increases societal vulnerabilities. This bidirectional instability increases the risk of internal conflict in both receiving and sending societies, while also significantly increasing the likelihood of regional and even international tensions.

Health kiosks, installed in public spaces, are becoming one of the most visible and controversial faces of digital health. Spanning from shopping malls and airports to university campuses and rural settlements, these stations offer a wide range of services, from simple screenings like blood pressure measurement to remote physician consultations and AI-powered preliminary assessments. This proliferation, which has accelerated in the last two years, is not just a technological innovation; it also raises deeper questions, such as access to healthcare, inequalities, and the transformation of labor.

Today, a person approaching a health kiosk often begins with self-measurement: basic indicators such as blood pressure, pulse, oxygen saturation, weight, and body mass index are recorded within a few minutes. Some advanced models also include electrocardiography (ECG) or risk scores. These measurements are often pre-screened by AI-powered software, providing the user with understandable feedback, such as “Normal,” “follow-up recommended,” or “consult a specialist.” Here, artificial intelligence is positioned not as a replacement for the physician, but as a layer that accelerates and guides the process.

The second approach is telehealth kiosks, called “clinic-in-a-box.” In these booths, users can have a video consultation with a clinician remotely. Thanks to sensors and cameras inside the booth, the physician can see the measurements in real time. OnMed kiosks placed in airports in the US are among the best-known examples of this model. The aim is to provide fast, low-threshold health contact, especially in areas with dense and transient populations.

In Germany, however, the approach is somewhat different. Here, the “Gesundheitskiosk” concept is more focused on multilingual counseling, preventive health information, and referral to the health system. In federal and state-level pilots, the goal is to facilitate access to the system, especially for immigrants and groups with low health literacy. Guidance and bridging functions are prioritized over clinical diagnosis.

Projections for the next three to five years show that these kiosks will become even more “standardized.” Basic vital signs will become almost a default feature; AI will focus on background functions such as triage, risk flagging, appointment scheduling, and documentation. Regulation is expected to be the decisive factor. In the European Union, the intersection of medical device legislation and AI regulations will force manufacturers into a stricter framework regarding quality, traceability, and accountability. This will increase the tension between “rapid deployment” and “secure integration.”

The critical question here is: Are health kiosks truly a solution that increases accessibility, or a new layer of inequality? Will groups with high digital literacy and less concern for privacy benefit more easily from these systems, while the elderly or those hesitant about technology will be left out? Who will be held responsible if there is an error in AI-powered triage? Is the kiosk a support that closes the physician shortage, or a tool that makes healthcare labor even more fragmented and precarious?

The issue is not just about “a new device”; The public nature of healthcare, points of contact for migrants within the system, data privacy, and how academic knowledge is transferred to the field are all relevant issues. Health kiosks, when designed correctly, can be a low-threshold and inclusive entry point; when poorly designed, they can become a symbol of silently deepening inequalities. Therefore, the issue hinges not so much on “will it happen?” but rather on “how will it happen and for whom?”

https://www.nature.com/articles/s43856-025-00738-5

https://link.springer.com/article/10.1186/s12872-023-03701-1

https://innovationsfonds.g-ba.de/downloads/beschluss-dokumente/140/2022-02-16_INVEST_Billstedt.Horn_Evaluationsbericht.pdf

We are academics.

We bear the responsibility of knowledge, critical thinking, and ethical consistency.

Therefore, our stance against violence cannot change according to identity, geography, or power.

We unequivocally and unconditionally condemn the attack targeting the Jewish community in Australia.

Antisemitism, with its historical weight and current manifestations, is a hate crime against human dignity.

It cannot be justified, regardless of where or by whom it is committed.

However, academic conscience requires us to look not only at individual incidents but also at the ongoing structural violence.

Today, Israel’s genocidal policies in Palestine have transformed into a process of destruction where civilians are systematically targeted,

living spaces are destroyed, knowledge, education, and academic existence are made impossible.

Stating this truth is not antisemitism.

On the contrary, it is about standing against approaches that instrumentalize the moral power of antisemitism to cover up state violence.

Academic solidarity is not about selective empathy.

It is not about ignoring the suffering of the powerless and remaining silent in the face of the powerful. It is certainly not about remaining silent by saying, “It’s wrong, but now is not the time.”

We know that:

Hate crimes must be condemned.

Genocide and policies of mass punishment must be stopped.

Academics, students, and intellectual life cannot be targeted in any geographical area.

Silence is not neutrality.

Silence is often taking a stand alongside the existing injustice.

Therefore, as Academic Solidarity, we declare:

We are against all forms of antisemitism.

We are against all forms of racism.

And we are against systematic, persistent, and destructive violence perpetrated by states.

Our solidarity is based not on identities, but on human dignity.

2025 is shaping up to be a year in which measles is on the rise globally. This disease, which has been kept under control for years thanks to effective vaccination programs, is now causing serious outbreaks in many countries. In the United States, in particular, measles cases have reached their highest level in the last thirty years. According to CDC data, nearly two thousand confirmed cases have been reported nationwide by 2025, and this increase jeopardizes the “measles elimination” status achieved in 2000 (https://www.cdc.gov/measles/data-research/index.html). The most striking increase is in South Carolina; nearly 126 cases have been detected in the state, and hundreds have been quarantined due to contact. Small school-based outbreaks are also continuing in Utah, Arizona, and many other states. The majority of this increase in case numbers is due to unvaccinated or incompletely vaccinated individuals (https://www.reuters.com/business/healthcare-pharmaceuticals/south-carolina-measles-cases-rise-126-amid-accelerating-outbreak-2025-12-12/).

A similar picture emerges in Europe. According to the World Health Organization and UNICEF, the European Region reported more than 127,000 measles cases in 2024, the highest level recorded since 1997. The increase continued throughout 2025; the European Centre for Disease Prevention and Control (ECDC) reported over 9,600 confirmed cases in EU/EEA countries from November 2024 to October 2025. Countries such as Belgium, Italy, and Spain stand out with hundreds of cases, while many countries are seeing an acceleration of the spread among school-aged children. Experts emphasize that disruptions to routine vaccinations and reduced access to healthcare during the COVID-19 pandemic negatively impacted immunity rates across the continent (https://www.who.int/europe/news/item/13-03-2025-european-region-reports-highest-number-of-measles-cases-in-more-than-25-years—unicef–who-europe).

A similar risk exists in other parts of the Americas. According to PAHO reports, by mid-2025, more than 10,000 cases and dozens of deaths were recorded in ten countries in Latin America. Although measles remains a completely preventable disease globally, WHO reports indicate that approximately 10 million people contracted measles in 2023. This picture clearly shows that routine immunization rates in some regions remain below the critical threshold (https://www.paho.org/en/news/15-8-2025-ten-countries-americas-report-measles-outbreaks-2025).

The main reason measles can spread so quickly is that the virus is highly contagious. Approximately 90% of unvaccinated individuals who share an environment with an infected person can contract the disease. In contrast, two doses of the MMR vaccine provide over 97% protection. Having at least 95% of the population vaccinated is critical for controlling measles at the community level. However, during the pandemic, disruptions in vaccinations, vaccine hesitancy, and the strengthening of misinformation campaigns have caused immunity rates to fall below this threshold in many countries.

Increased global human mobility and forced migration may also be a reason for the increase in measles cases. Experts emphasize that migration alone is not the cause of outbreaks; However, disruptions to vaccination programs in conflict zones and low-income countries, inequalities in access to healthcare, and the harsh conditions of migration journeys can lead to vaccine shortages in some communities. This makes it even more critical to include all migrants in early vaccination programs, especially in destination countries. However, it is noted that the determining factor in the recent measles surges in Europe and America is not migration, but rather declining vaccination rates and increasing vaccine hesitancy within the local population. Therefore, it is necessary to address the relationship between migration and health within a framework that emphasizes the importance of equitable access to public health and strong immunization programs, avoiding one-sided explanations.

All this data shows that measles is not only a medical problem but also a global public health issue. The disease can have serious consequences, including pneumonia, encephalitis, and death, especially in young children. Despite having an effective, inexpensive, and safe vaccine, we are still seeing major outbreaks in many parts of the world. Public health experts state that rapid and comprehensive campaigns to increase vaccination coverage are necessary, that the public is supported with scientific information, and that active combat against anti-vaccine misinformation is essential.

This measles wave in 2025 reminds us once again how vital immunization programs are and the importance of strengthening routine health services in the post-pandemic world. The fact that we are still talking about outbreaks today for a disease for which vaccines are effective clearly reveals how fragile global health systems can become and the importance of science-based public health policies.

Approximately 90 people from various countries and healthcare disciplines attended the online seminar “AI in Healthcare,” organized by Akademische Solidarität e.V. on December 1, 2025. The seminar speaker was Dr. Tolga Ensari, a faculty member in the Department of Computer Science at Arkansas Tech University.

At the beginning of his speech, Dr. Ensari summarized the development of artificial intelligence and its fundamental mathematical and statistical infrastructure, emphasizing that AI is not just about programming; it possesses a broad knowledge base, ranging from differential equations and optimization to graph theory and probability models.

In the main part of the speech, Dr. Ensari discussed the current applications of artificial intelligence in healthcare:

• AI-assisted diagnostic systems in medical imaging (CT, MRI, pathology) have become widely used in areas ranging from cancer detection to image interpretation.
• Clinical decision support software (e.g., IBM Watson Health) is accelerating physicians’ diagnostic and treatment processes.
• The transformation of text, genome data, or sensor outputs into AI algorithms was explained with examples.
• Emerging areas included digital mental health, biological computers, the concept of next-generation “actual intelligence” working with nerve cells, and biologically inspired models such as spiking neural networks.

Dr. Ensari also explained with concrete examples that artificial intelligence, like humans, can “make mistakes.” Tesla cars in Australia failing to recognize kangaroos and stopping, or image classifiers confusing the difference between a muffin and a puppy, were shared as examples demonstrating that AI systems should not be considered absolute accuracy.

The seminar covered ethical principles and upcoming legal regulations extensively. Within this framework, data privacy, fairness and transparency, human control, manageability, and traceability were defined as fundamental conditions for AI systems.

The seminar also shared examples from the comprehensive “technology constitution” studies that have been underway in the US and EU for five years. The legal liability of physicians when using AI, shared responsibility in potential malpractice lawsuits, and the direction of future regulations were discussed. Dr. Ensari also answered the questions of the participants:

• It is now common for patients to consult tools like ChatGPT before seeing a doctor; however, it may be ethically more appropriate for the doctor to conduct the conversation with the AI ​​in front of the patient and in a transparent manner.
• It was emphasized that if the AI makes an incorrect decision, the software cannot be held solely responsible.
• It was stated that AI systems capable of meta-analysis are becoming increasingly powerful and could soon completely transform scientific research processes.
• It was stated that the use of AI for casual conversations on mental health-related topics is harmful, but licensed “Digital Mental Health” systems could be beneficial.

At the end of the seminar, one participant discussed the practical benefits of AI-powered software like Tandem Health, used in primary care in Sweden:

• Automatically transcribing physician-patient conversations and converting them into medical notes,
• Creating correspondence such as reports, referrals, and sick leave documents in seconds,
• Automating drug interactions and guideline checks.

Another participant discussed Heidi Health, a transcription and decision-support tool used in Germany, emphasizing the importance of integrating this software into local systems, particularly due to data security requirements.

“Time to Get Out of Your Comfort Zone”

At the end of the program, when one of the moderators remarked that the seminar had “pushed him out of his comfort zone and into a fear-learning-development cycle,” Dr. Ensari responded with the following words:

“Being afraid is good. Because it’s the beginning of learning and development.”

This comprehensive seminar, organized by Akademische Solidarität e.V., demonstrated how AI is transforming healthcare, from technical, ethical, and practical perspectives. The intense interest and active contributions of the participants demonstrated once again that healthcare professionals are ready for this change, but that the road is still very early.

This week’s decision by the UK’s national screening committee not to recommend prostate cancer screening for the general population has sparked a new debate in the international medical community. The rationale behind the decision is familiar: While PSA (prostate-specific antigen) testing can save some lives, it also carries the risk of unnecessary diagnosis and unnecessary treatment. Furthermore, recent evidence suggests that this balance remains delicate (https://www.ft.com/content/9065a8d8-8bfb-40e7-a64d-a326275a00e8).

The 23-year results of the European ERSPC randomized trial, one of the most comprehensive screening data, are as follows: The relative likelihood of dying from prostate cancer is approximately 13 percent lower in the group receiving regular PSA screening. However, this “relative” difference translates to a very small real-world gain: over a 23-year follow-up, only 2 or 3 out of every 1,000 men screened appear to have prevented prostate cancer deaths. In contrast, the rate of prostate cancer diagnosis increases by 30 percent in the screened group, meaning many men are labeled “diseased” for an indolent tumor that will never cause problems. This can mean unnecessary biopsies, unnecessary surgeries, and complications that diminish quality of life (https://pubmed.ncbi.nlm.nih.gov/41160819/).

One of the most important questions in prostate cancer screening remains the accuracy of the PSA test, a fundamental tool. The false-positive rate of PSA is quite high; for example, approximately 70 percent of men with elevated PSA levels do not have cancer detected upon biopsy. Prostate enlargement, infections, and even recent sexual intercourse can all elevate PSA levels. This translates to a significant burden of anxiety, unnecessary biopsies, and sometimes unnecessary treatments. On the other hand, the false-negative nature of PSA cannot be ignored; approximately 15% of prostate cancers can be missed even when the PSA is normal. This rate maintains the concern about aggressive tumors (https://bmjoncology.bmj.com/content/2/1/e000039).

Another prominent issue in current discussions is which method should be preferred if screening is to be performed. In the past, digital rectal examination (DRE) was routinely recommended for prostate cancer screening. However, recent studies have shown that DRE has very low sensitivity when used for screening. Most current guidelines emphasize that DRE should be considered as a complementary physical examination tool in men with specific complaints or high PSA levels, rather than as a screening test. If a screening decision is to be made, the PSA test is the preferred method due to its higher sensitivity. However, it should be remembered that PSA alone does not provide definitive results and carries the risk of both false-positive and false-negative results. Many countries are now adopting a more selective approach, rather than recommending screening for the entire population.

The USPSTF recommendations recommend screening men aged 55–69 only after a comprehensive consultation between physician and patient, with a clear discussion of the potential benefits and harms (https://www.uspreventiveservicestaskforce.org/uspstf/recommendation/prostate-cancer-screening). It emphasizes that the harms of screening in men with limited life expectancy or over 70 outweigh the benefits. The UK’s recent decision, however, rejects universal screening but supports biennial screening only for men with BRCA1/2 mutations, as prostate cancer can progress earlier and more aggressively in this group. Germany is also ending the systematic “rectal examination + automated biopsy” era; the German guideline recommends a more personalized strategy (based on risk, age, PSA value, demand, etc.) based on PSA. However, this strategy is still not “mandatory for everyone” or a “mass screening program” (https://register.awmf.org/assets/guidelines/043-022OLl_S3_Prostatakarzinom_2025-08.pdf).

Recent medical advances indicate that multiparametric prostate MRI, in addition to PSA, is becoming an increasingly powerful tool. MRI has the potential to reduce unnecessary biopsies and detect clinically significant cancers more accurately. However, the applicability, cost-effectiveness, and ideal age ranges for MRI-based screening in the general population are not yet fully understood (https://bmjopen.bmj.com/content/12/11/e059482).

All these data clearly demonstrate that prostate cancer screening is not a simple “have it or not have it” decision. Each man’s risk profile, life expectancy, family history, and personal preferences are different. Good communication between physician and patient requires honest discussion of the potential benefits of screening, as well as the physical and psychological harms that can result from overdiagnosis, overtreatment, and testing errors.

Ultimately, current evidence does not support mandatory or automatic screening in the general population. However, Screening should be carefully considered for healthy men aged 50–70, especially those with a family history or in high-risk groups. The sole purpose of screening is not early detection; it also empowers men to take an active role in their own health decisions. In medical practice, the goal is not more testing, but more information and less harm. Current debates surrounding prostate cancer screening call us to this very point.

Retracted papers are among the most drastic and visible warning signs used to protect research integrity in the scientific literature. However, a new study published in Retraction Watch on November 19, 2025, shows that the rapidly proliferating AI chatbots are having a particularly difficult time recognizing these critical warning signs. The researchers warn that academics, particularly those relying on ChatGPT and similar tools, risk serious errors if they use the responses of these models as an “automatic truth filter” (https://retractionwatch.com/2025/11/19/ai-unreliable-identifying-retracted-research-papers-study/).

Konradin Metze and his team at the State University of Campinas, who conducted the study, designed a relatively simple experiment. They presented a list of publications by Joachim Boldt, known for his major scientific fraud scandal in anesthesiology, to 21 different AIs. The list included the most cited retracted Boldt articles, the most cited Boldt publications that had not been retracted, as well as articles written by other authors with the last name Boldt. For each of the 132 references, the bots were asked a single question: Was this article retracted or not?

The results were striking. Most of the chatbots correctly identified less than half of the retracted articles. Not only did they miss them, but they also incorrectly marked a significant portion of the unretracted articles as retracted. This represents a serious weakness in both sensitivity and specificity: the AI provides false assurances and casts unnecessary doubt on well-established articles.

When the research team repeated part of the experiment three months later, they encountered an even more striking pattern. In the first round, the bots generally used definitive statements, but in the second round, they began using vague and evasive phrases, such as “possibly retracted” or “requires further review.” The researchers interpret this shift as the models oscillating between “offering false certainty” and “trying to save themselves with vague statements.”

The Retraction Watch report also cites another recent study by Mike Thelwall of the University of Sheffield. Thelwall had ChatGPT evaluate 217 retracted or seriously questioned papers 6,510 times. In none of these thousands of responses did ChatGPT indicate that the paper had been retracted, raise questions about it, or contain scientific issues. On the contrary, it even praised some retracted papers as “high-quality work.” This demonstrates that AI not only misses retraction information but can also glorify and reproduce erroneous or false scientific findings (https://sheffield.ac.uk/ijc/news/new-research-suggests-chatgpt-ignores-article-retractions-and-errors-when-used-inform-literature?utm_source=chatgpt.com).

The problem isn’t just recognition. Another study published in the Journal of Advanced Research revealed that chatbots use retracted articles as sources in their responses. This means that AI can now recirculate information that was previously considered obsolete in the scientific literature. As more and more people in academia use tools like ChatGPT to quickly summarize, develop research ideas, or master the literature, the risk of recirculating retracted information becomes increasingly significant.

Sociologist of science Serge Horbach calls these developments a “clear warning”: LLM models are not suitable tools for weeding out retracted articles. The training data for AI models is fed by a system that is both historically lagging and where retraction information is published in a fragmented manner. Information about an article’s retraction may be visible only on the journal page, only in PubMed, or only in the Retraction Watch database. Scanning this fragmented structure with security and accuracy is far beyond the technical capabilities of today’s chatbots.

For Academic Solidarity, these findings hold particular significance for academics in exile or working in precarious circumstances. In situations where access to research infrastructure is limited, tools like ChatGPT offer attractive speed and convenience. However, this convenience carries the risk of unnoticed reproduction of studies based on retracted or inaccurate information. This risk can be even more severe for researchers working in political, legal, or human rights fields; misinformation cannot only be a scientific error but also open the door to political manipulation.

This situation doesn’t necessarily mean AI should be completely excluded from research processes; however, it does highlight a critical limitation: ChatGPT and similar models are not reliable filters for detecting retracted literature.

This week, the Academic Solidarity Association is examining the relationship between academia and the real sector. While the critical role of universities in innovation and development is becoming increasingly stronger, collaboration between the two sides remains significantly below potential. This is a common concern not only in Turkey but also in Germany and other OECD countries. The OECD’s latest report on university-industry collaboration indicates that collaboration across countries falls short of expectations, citing bureaucracy, financing difficulties, and the different working rhythms of the two institutions (OECD 2023, https://www.oecd.org/sti/university-industry-collaboration.htm).

The criticism of academia as “working in glass palaces” is a frequent topic of public discussion. Researchers analyzing the academic world have noted that the pressure to publish, high teaching loads, and administrative duties consume a significant portion of academics’ time. This situation makes collaboration with external sectors difficult (Altbach 2015, https://doi.org/10.6017/ihe.2015.79.5837 ). Similarly, the European Commission emphasizes that there is a systematic disconnect between academia and industry due to “different motivations and lack of communication” (European Commission 2021, https://place-based-innovation.ec.europa.eu/publications/higher-education-smart-specialisation-handbook_en).

Meanwhile, the effectiveness of universities in the real sector is also a matter of debate. It is stated that businesses in Turkey generally see universities as institutions that provide qualified graduates, and that joint projects focused on R&D and innovation are still limited (https://www.yok.gov.tr/documents/documents/68c01f9a0dc63.pdf). While the situation in Germany is better, it is not entirely different. The “Transferindikator Deutschland” report, prepared by the Stifterverband and CHE, shows that even within German universities, sectoral collaboration remains below potential (Stifterverband & CHE 2022, https://www.stifterverband.org/transferkompass).  

One of the key reasons why sector expectations from universities are not being met is a lack of practical experience. The World Economic Forum’s “Future of Jobs” report states that university graduates in many countries struggle to meet the needs of the business world in terms of practical skills (WEF 2020, https://www.weforum.org/reports/the-future-of-jobs-report-2020).

Another factor weakening the academia-industry relationship is mutual distrust. A comprehensive review published in the journal Research Policy reveals that sector perceives academia as “slow and abstract,” while academia perceives sector as “impatient and commercially oriented.” The underlying causes of this situation lie in a lack of communication, differing motivations, and the weakness of intermediary institutions (Perkmann et al. 2013, https://doi.org/10.1016/j.respol.2012.09.007).

Successful examples from around the world prove that this disconnect can be overcome. The fact that approximately 70% of the Fraunhofer Institutes’ revenues come from private sector projects and the institution’s global success in applied research provides a strong model (Fraunhofer Jahresbericht 2023, https://www.fraunhofer.de/en/annual-report.html). The university-startup ecosystem that has developed around Stanford and MIT in the US enables the rapid commercialization of academic research (Roberts 2019, https://www.nowpublishers.com/article/Details/ENT-093). Local innovation programs implemented through the municipality-university-industry triangle in countries like Sweden and the Netherlands offer good examples of how collaboration can be institutionalized.

So, how can academia and the corporate sector work more effectively? According to an OECD analysis of collaboration policies, joint funding mechanisms, tax incentives, and strengthening technology transfer offices are among the critical steps for sustainable collaboration (OECD 2019, https://www.oecd.org/sti/university-industry-collaboration-policies.htm). Industrial doctoral programs, increasingly widespread in Europe, academic consulting, and developing a collaborative project culture with industry are also frequently recommended strategies.

While there is a visible distance between academia and the corporate sector, it’s clear that there is significant untapped potential. In areas like digitalization, artificial intelligence, sustainability, and global competition, both sides need each other much more than before. With the right bridge mechanisms, strong intermediary structures, and political support, it’s possible to create true synergy between universities and the business world. This collaboration is crucial not only for economic growth but also for social development and scientific advancement.

Students took the stage at Istanbul Technical University’s recent graduation ceremony to protest the rector’s speech. Families in the stands responded with applause. This unified reaction at the ceremony was not merely a momentary outrage, but a manifestation of a cumulative, collective unease about university administration and academic freedoms in Turkey (https://onedio.com/haber/itu-mezuniyet-toreninde-rektorun-konusmasini-protesto-eden-ogrencilere-aileler-alkislarla-destek-verdi-1303148).

Universities in Turkey have shifted to a significantly centralized governance model over the last decade. Regulations enacted after 2016 eliminated rector elections entirely, and all rectors were appointed by the President. Boğaziçi University, METU, Istanbul University, Marmara University, and many other established universities have experienced ruptures in both their academic and cultural identities due to appointments made against the will of their faculty members. These appointments are not merely administrative decisions, but are considered an erosion of the principle of autonomy that universities have historically upheld.

This situation is not unique to universities. We observe similar tensions extending to high schools, as seen at the graduation ceremony of Ankara Science High School. Students protested the school principal, believing they were being stripped of their right to shape their educational environment. This perception of diminished participation and voice in education is now felt not only among university students but also in younger age groups (https://ankahaber.net/haber/detay/ankara_fen_lisesi_ogrencileri__mezuniyet_toreninde_okul_mudurunu_protesto_etti_246956).

Doruk Dörücü’s protest, which involved tearing his diploma on stage at the Boğaziçi University graduation ceremony (https://www.dw.com/tr/i%CC%87mamo%C4%9Flu-protestosu-diplomas%C4%B1n%C4%B1-y%C4%B1rtan-doruk-d%C3%B6r%C3%BCc%C3%BC-serbest/a-73153618), has become etched in memory as one of the most symbolic examples of this transformation. While the words “This despotic regime’s diploma is invalid” may seem like an individual statement, they were an expression of the feeling that the university had ceased to be a “home” for students where they could express themselves. Tearing up the diploma was a protest not only against the educational institution itself, but also against the cultural meanings it was believed to represent.

These developments are not unique to Turkey. During the same period, universities in the US (https://www.independent.co.uk/news/harvard-university-donald-trump-university-of-kentucky-education-department-phoenix-b2742772.html) and elsewhere (https://www.belfasttelegraph.co.uk/news/northern-ireland/ulster-university-accused-of-censorship-after-removing-palestinian-flag-footage-from-graduation-video/a231930227.html) are canceling graduation ceremonies due to campus protests and political pressure, banning symbols and limiting students’ right to speak. The politicization of academia and the government’s reaction to taking precautions against free expression are exhibiting a similar trend on a global scale. Forms of oppression, like the international circulation of knowledge, no longer recognize borders.

The applause rising from the stands at ITU signaled a significant turning point. This time, it wasn’t just the students speaking; Families also demonstrated their rejection of the pressure placed on the stage. This support reminded them that academic freedom is not merely an internal debate among academics or students, but a shared value that shapes the future of society.

Graduation ceremonies are rituals that demonstrate how a university defines itself. If these rituals are forced into silence, the intellectual space of the university is also shrinking. What happens on the graduation stage in Turkey today is not just a ceremony, but a discussion about shaping public reason, cultivating a culture of criticism, and envisioning a social future.

And perhaps precisely for this reason, that applause at ITU is more than a protest. The most fundamental expression of academic freedom is this: Knowledge is only meaningful when it is freely produced. The way out of this impasse is for the public to collectively respond to the pressures placed on the academy and hold politicians accountable. The university is not isolated from society; where society remains silent, the university is silenced.