/  STEAM Careers   /  AI in STEM – Revolutionary Change across the Life Sciences

AI in STEM – Revolutionary Change across the Life Sciences

Every day, we see the impact of technology everywhere. No matter how we feel about this technological revolution, we can expect that it will only continue to expand and become an even more significant component of our lives.

A key area that continues to progress significantly is artificial intelligence. The innovation in this field has grown so quickly and substantially, that AI is already expanding across other areas, including science.

In our last article, we kicked off a short series exploring the role of AI in STEM by discussing how AI is revolutionizing scientific research. This time, we’ll be looking in-depth into the numerous applications of this technology in various life sciences.


While we’ve touched on genetics when we discussed the case of population genomics in our previous article, that is only the tip of a much larger iceberg. This may be one of the life sciences that most brilliantly exemplifies the immense power AI holds as an aid for researchers.

For one, AI can be extremely useful to analyze and interpret the massive volume of data generated in this field. In particular, AI’s capacity to discover patterns in data can help, for example, to identify genetic variants that cause certain diseases.

But that is only one part of the wider picture, as genetic information doesn’t fully account for the characteristics we observe. There’s an increased focus today on studying each of the different factors involved and understanding how they connect – this is what’s referred to as a multi-omics approach.

Essentially, when considering any condition or characteristic, we should not only look at the genomic data but also information about other aspects like the microbiome and the epigenome. To learn more about multi-omics, read this short explanatory article.

AI can help us connect these various categories of data, highlighting relevant patterns and making predictions of how different factors act together.

One example of this is Enformer, an AI program created by DeepMind and Calico Life Sciences, that analyzes DNA sequences and from that predicts how genes are likely to be expressed. You can access DeepMind’s article, discussing this program in depth, through this link.

Finally, AI is also being used to enhance gene editing tools like CRISPR, increasing their accuracy and therefore diminishing the risk of off-target effects. This is a very exciting development with the potential to significantly impact areas like drug discovery.

To learn more about this particular topic, we’d recommend reading the following research article, perhaps with the help of an AI tool like Explainpaper.


Despite all of our progress, the brain is such a complex organ that there is still much that is unknown about its functioning. Understanding the dynamics of the thousands of neurons in our brains, as well as how these can change and cause different disorders, is an extremely arduous task.

Once again, AI can help overcome some of the challenges in this field, in particular through its capacity to find patterns and build complex simulations that often factor in components that humans may not even perceive.

One specific application of AI that could lead to very significant changes in neuroscience is the creation of individualized representations. This is something that is only starting to be explored, but if achieved could significantly improve our capacity to care for people with brain-related conditions, including mental disorders.

The goal would be to use each person’s data to create simulations of their brain that may enable clinicians to put together individual treatment plans with more ease and accuracy.

Hopefully, as AI continues to evolve and expand into other areas, we’ll see more applications within neuroscience that help advance this field.

You can find out more details about the current applications of artificial intelligence in neuroscience in this article.


One of the major areas for the application of AI right now is imaging. In particular, medical imaging has increasingly become a priority when considering the use of AI, not only because of its importance for the diagnosis and treatment of several conditions but also due to the volume of work that is associated with it.

Medical imaging techniques, like magnetic resonance imaging (MRI) or X-ray computed tomography (CT), produce complex images that need to be carefully interpreted by physicians. However, AI can be trained to make assessments through these same images.

Essentially, AI can analyze them and find specific patterns and features in the data that can be used as biomarkers. This is a process referred to as data mining.

Image by Freepik

For example, Google collaborated with several institutions to develop an AI model that could be used to analyze digital mammography images for breast cancer screening. As discussed in an article by Google, diagnosing breast cancer at an early stage is still a tremendous challenge, even when these imaging techniques are used, given the difficulty in reading the images produces.

Even experts can make inaccurate assessments, but to make matters worse, there is a shortage of professionals available to assess digital mammography images. Nevertheless, the AI model created by Google can detect the presence of cancerous cells in these images with even more accuracy than experts, and thus could become a vital asset for breast cancer screening.

Another recent example of this type of application of AI, also developed by Google, was related to the diagnosis of tuberculosis. Due to the symptoms associated with this disease, it can be hard to diagnose it. While imaging techniques like chest X-rays can help identify this disease, there is again a lack of professionals capable of interpreting these images.

In response, Google developed an AI system that can help identify tuberculosis through chest X-rays with great accuracy, facilitating the diagnosis process. For more details, you can read this article.

These are only two examples and, as such, don’t represent the full scope of potential applications of AI in medical imaging.

AI may not just be useful in the diagnosis of diseases but also in assessing patients’ risk of developing specific conditions and determining an individual’s response to treatment.

Further, other areas utilize imaging techniques in which AI can also be used. Most often people tend to only focus on the application of AI in medical imaging, but similar systems can also be used, for example, in preclinical research trials.

A major advantage of this would be in facilitating the analysis of experimental treatments, by enhancing our capacity to understand how they work in vivo and the effects they can have on living organisms.

For example, InVivo Analytics created an AI-based tool that helps researchers determine the distribution of bioluminescent reporters within small research animals, such as mice. To put it simply, these reporters are molecules that emit light. As such, they are used in research to monitor the distribution of another molecule or even a microorganism that it can attach to.

In this specific study carried out by InVivo Analytics, researchers used a technique called bioluminescence imaging (BLi) to monitor a bacterial infection. This was selected in particular to test this new AI tool given that, due to the nature of this technique, automatic data analysis isn’t typically possible. However, the study showed that this tool could help to overcome this issue, facilitating the entire process and reducing errors.


Image by storyset on Freepik

AI is also being explored as a tool for application in the field of microbiology, not only within research but also in a clinical setting. However, as can be expected, there is a greater gap between innovation and the application of this technology in the medical field.

As indicated by this article, artificial intelligence is already being used to analyze a wide range of data types. These include genomic and metagenomic sequencing data, protein structures, transcriptome data, microscopy and culture images, and even clinical information.

In particular, from what we’ve previously discussed, is easy to see the benefit of using AI to analyze datasets that often are very extensive. A clear example of this is metagenomics, which looks into the nucleotide sequences found in a given bulk sample – such as a sample of soil – from the various organisms in that environment. In these cases, AI is essential given the volume and complexity of the data obtained.

It’s only natural that in a field like microbiology, which has seen incredible technological advances in the last decades, we need to employ revolutionary tools such as AI.

Research already shows that there is a wide range of applications for AI in microbiology, from detection and identification of microorganisms, to testing of their susceptibility to a series of antimicrobial drugs (referred to as susceptibility testing), to diagnosis and prediction of a patient’s outcome.

Most of the applications explored so far have focused on automating repetitive tasks and, as such, a microbiologist is still required to validate the results obtained at the end. Even so, using AI can significantly reduce the time and cost expended on these tasks.

Most importantly, experts believe that the use of AI in microbiology will drive us to create innovative tools and techniques with far-reaching applications.

A good example of this was the creation of a smartphone application using AI to detect malaria parasites in thick blood smears. The diagnosis of malaria typically relies on microscopy analysis of blood smears; however, this technique is very time-consuming and must be carried out by an expert to lead to an accurate diagnosis.

That led researchers to develop an app that could automate this process not only to make it faster and more efficient but also more accessible to locations that lack the resources typically required for diagnosis.

You can read more about this specific topic through this article.

Healthcare and Personalized Medicine

So far, we’ve already discussed a few of the ways AI is being used to enhance medical practice, from medical imaging to the diagnosis of infectious diseases. Nevertheless, there are many other ways in which this type of technology is being used in healthcare.

These extend through several areas, from systems to organize and manage health records and software that helps physicians decide on a patient’s diagnosis and treatment, all the way to surgery robots and nanorobots used (amongst other things) for targeted drug delivery and microsurgery.

For example, researchers in Switzerland are currently developing a project using nanorobots, guided by ultrasound, to deliver drugs to specifically target cancer cells.

This is only one of several projects highlighted by the European Commission in this area, so if you’re interested in learning more, this article is a great place to start. You can also read the following review for a slightly more technical overview of nanorobots.

Another form of application of tech in healthcare, which most of us are more familiar with, is health-related apps and biosensors.

These can help us in a series of manners, whether it be managing the user’s health and promoting changes in their behavior, collecting data for research, or even direct application in medical practice.

We can already see this in a few different areas. Today, many of us have devices like smartwatches that help us monitor parameters related to our health including our level of physical exercise throughout the day, our sleep patterns, and our heart rate.

These can not only help us manage some aspects of our health, but also modify our behaviors. For example, the popular app Headspace takes advantage of machine learning to personalize its features as the user explores the app, enhancing their experience and promoting the continued use of their mental health services.

You can find out more about this through this article posted by the company.

Of course, the same principles can be applied to other conditions, perhaps with a more significant connection to formal care. For instance, Google was involved in a partnership to help develop an AI system that helped monitor the health of people with cardiovascular diseases, giving them tailored advice throughout the day in accordance with their activity and medical history.

For more details, you can assess the full article on this topic here.

In the future, the data that is collected through these apps and devices could also then be applied to research, to increase our knowledge of the human body and of health and disease, so that we can make more discoveries that improve healthcare.

Particularly, leveraging AI in this process could significantly help analyze the large volumes of data collected, facilitating the process, and enhancing the research conducted. If you want to learn more about how AI is used in scientific research, you can read our previous article in this series.

Hopefully, then, these developments can lead us towards more holistic and personalized care, in which medical decisions are based not only on the patient’s immediate examination but also on their omics data, medical history, behavior, and environment.

As we know, people’s bodies don’t function the same and all of the factors described have an influence on our health, as well as our response to therapeutic interventions when we’re ill. For example, everyone can have distinct reactions to the same drug, even common ones like Tylenol (or paracetamol). Nevertheless, the guidelines for diagnosis and treatment are always based on the response of the average patient.

Precision medicine instead focuses on adapting these practices to each patient, so that instead of, for example, attempting treatments through trial and error, all factors are considered first so that the patient is given the ideal drug for them.

Of course, this is not yet possible given the gaps in our understanding, but AI holds great promise in helping to develop this prospect. To learn more, we’d recommend the following article.

Environmental Studies and Climate Change

One of the greatest goals in science today is to not only understand how individual organisms work and how they interact with each other, but also how their ecosystems function as a whole. With this knowledge, we would be much better equipped to protect the environment and all the beings that inhabit our planet.

For one, we would be able to monitor ecosystems and detect any changes and anomalies early on, allowing us to tackle any issues more effectively, perhaps even before they manifested. To some extent, this is already possible on a smaller scale, as we’ll explore later.

A deeper understanding of ecosystems and the living beings that live in them could also enable us to engineer life. While today we can modify certain bacteria and other simple organisms, in the future, we could eventually make changes at a wider scale. It could even be possible to control and improve entire ecosystems, for example, by creating more arable areas for cultivation or enhancing water quality.

For now, however, we need to focus on climate change, which threatens to destroy our planet and the creatures living on it, including humans.

In 2021, a group of scientists launched a very interesting AI-based project called “This Climate Does Not Exist” to illustrate the impact of climate change. Using a machine learning algorithm, images are originated that depict the effect of floods, wildfires, and smog in any location you wish.

This is only one of many ways AI is being used in the fight against climate change.

To start off, AI is helping us to understand the effects of climate change and how we can respond to these. It’s already extensively clear that no matter what mitigation efforts we employ now, there are some effects of climate change that we’ll still have to endure, simply because we failed to act earlier.

In some cases, all we can really do now is attempt to predict any major weather events, like wildfires or floods, and implement measures that help people respond to these, to avoid as many casualties and damage as possible.

AI is already being applied in this sense, as several companies and organizations are creating and implementing AI models not only to help predict these catastrophes but also to make effective decisions to tackle them.

For example, Google created an AI system that uses satellite images to identify wildfires in real time and help predict their course. This can be extremely helpful to the authorities combating these fires and evacuating people as necessary.  

Google also created a similar system to predict floods, their extent, and their effect. This information is shared through a platform called “FloodHub” so that those at risk and the local authorities can more effectively prepare ahead of floods.

However, not all damage can be avoided and so, AI can also analyze satellite images to assess the effects of a given extreme weather event, providing information that can help authorities identify priority areas for disaster relief. You can learn more about these and other similar projects from Google through this article.

Of course, the main goal concerning climate change is to find ways to stop it. To do this we not only need to understand what is happening and what factors are contributing to climate change (and how) but also what we can realistically do to fight this in the most effective way.

All of this revolves around the collection and analysis of large volumes of data focused on a series of variables, from greenhouse gas (GHG) emissions to the sea level or the ocean temperature. What is a wonderful tool to do this? Artificial Intelligence.

Google is using AI, amongst other things, to monitor air quality across the globe or to directly optimize traffic lights to avoid congestion and, thus pollution. In another case, Google partnered with Wadhwani AI in India to create an AI tool to help farmers identify pests and ways to treat them.

You can learn more about the multitude of special projects developed by Google, related to themes like fishing, GHG emissions, solar energy, or deforestation, through this link.

While each project can have its own individual impact, scientists and government officials oftentimes have to consider several factors and how they affect each other to make appropriate mitigation decisions. Both Google and Microsoft, for example, have developed platforms that try to facilitate this process, by combining data from different climate-related factors.

In Google’s case, they created the Environmental Insights Explorer (EIE), an online tool that compiles data related to carbon emissions. It shows the emissions sources in each city, including transportation and buildings’ emissions, and highlight strategies that may help reduce these. This link will take you to the EIE’s website where you can find out more information and experiment with this tool.

Microsoft on the other hand created the Planetary Computer, which uses AI to analyze and present data on multiple variables, such as biodiversity, air quality, or land use, on a map. This platform is already being used, for example, to monitor ecosystems and plan conservation strategies, as discussed on the platform’s website.

Overall, it’s abundantly clear that artificial intelligence holds great promise in the fight against climate change. It can help us, not only to respond to major weather events but also mitigate the negative changes that are impacting our planet.

For more in-depth information on this topic, we’d recommend reading the report created by BCG and AI for the Planet, “How AI Can Be a Powerful Tool in the Fight Against Climate Change”, as well as “Climate Change and AI – Recommendations for Government Action”, created by GPAI, Climate Change AI and the Centre for AI & Climate.

Ecology, Botany, and Zoology

So far, our discussion on climate change has mostly focused on its impact on humans and the spaces we generally occupy. But what about the other creatures that inhabit our planet?

Image by storyset on Freepik

Fauna and flora are both deeply affected by this issue, however we also have to consider the impact directly caused by human activity.

Particularly with the growth of the human population, we’ve not only intensified agriculture and livestock farming, but also invaded and destroyed habitats and, in some cases, continued unethical practices like poaching.

As can be expected, this has a significant effect on animals and plants. Not only are their behaviors changing, but species continue to become endangered and even extinct, leading to dramatic changes in the dynamics within various ecosystems.

Now more than ever, it’s essential to focus on conducting more research to help us understand how ecosystems are changing and how the fauna and flora are being affected so that we can construct effective conservation strategies to help protect biodiversity.

As described in this article by Nature, our ability to monitor and study ecosystems and the beings inhabiting them has been revolutionized by emerging technologies. We now have access to tools such as satellites that allow high-resolution remote sensing, camera traps, acoustic sensors, and tracking tags.

These are originating large volumes of data that we simply aren’t able to process and analyze effectively. However, we can employ AI to do this instead, given its learning and analysis capabilities.

The goal would be for AI to help us comprehend the changes that are occurring, how we can mitigate these – potentially by altering our practices – and how we can keep monitoring and protecting biodiversity.

A prime example of this relates to an open-source platform designed by Marxan in collaboration with Microsoft and The Nature Conservancy, which allows users to make informed conservation decisions.

As discussed in this article from The Nature Conservancy, this is a spatial conservation planning platform. This essentially means that users can visualize and analyze global ecosystems data (relating to factors such as biodiversity distribution or land use) and use it to plan their conservation projects more efficiently while ensuring that people are not affected.

Of course, large-scale conservation projects aren’t the only area in which AI is being applied. There are several smaller AI-centered projects, focused on very particular issues in Botany and Zoology. These contribute, nonetheless, to biodiversity preservation given the interconnection between all of these fields.

One such example was the creation of two similar apps PlantMD and Nuru, which use machine learning to help identify diseased plants. In the case of PlantVillage’s app Nuru, AI was trained to specifically assess cassava, an important food source in Africa, indicating to users what disease was affecting their plant and how they could best manage it. For more details on this project, you can read this article.

AI is also being used to identify and characterize plants. As genotyping technologies progressed, genetic sequencing of specimens became faster and more accurate. However, our capacity to catalog plants and associate this new genetic data with information about plants’ phenotypes remained essentially the same.

As such, AI is being used to process the large volume of plant images and data, as seen with the ARADEEPOPSIS program, to analyze and assess differences between Arabidopsis plant specimens. As described in greater detail in this research article, artificial intelligence programs can be used to identify plant specimens, analyze their biological features and reproductive characteristics, and assess the distribution of certain species.

Why is this so important? Besides the obvious benefits of expanding our knowledge in the field of botany, this can be extremely beneficial to humans in two main areas – healthcare and bioengineering.

For one, understanding the genetic basis behind plant traits can be used to manipulate crops, improving their characteristics and yield, which is especially valuable to us as consumers, by enhancing the quality of food, improving its availability, and potentially reducing its cost.

In terms of our health, several human congenital diseases can be studied using plant models and, as such, gaining more knowledge about their genetics can help us better understand these conditions.

In relation to zoology, artificial intelligence can be used to identify and count animals, monitor them, and assess their behavior, which is not only extremely useful from a research point of view but also when considering wildlife preservation.

Recently, the Mathis Lab revealed a new AI-based software called DeepLabCut that can help monitor animals’ movements. It can assess whole-body movements, as well as more specific behaviors, such as hand articulations or eye movements.

As described by Nature Protocols, this software can be applied in various ways, from tracking movements to predicting poses, assessing an animal’s movement in 3D, or even studying precise behavior patterns associated, for instance, with decision-making. Overall, it seems like this tool could be very useful to monitor and study animal behavior.

Furthermore, DeepMind is utilizing AI to detect and identify animals through images from motion-sensitive cameras in the Serengeti National Park and the Grumeti Reserve in Tanzania. This project allows ecologists to monitor the animals in these protected areas, facilitating conservation efforts. You can find out more information about this project here.

Similarly, the Ruaha Carnivore Project is helping to protect large carnivores in the Ruaha landscape. In this case, AI was applied to camera traps to automatically detect and capture photos of these animals. However, this was not simply meant as a method for monitoring wildlife. Rather, it is used as a strategy to combat one of the biggest threats to these animals – the local community.

Locals in this area often entered into conflict with these animals and killed them, not only because they likely saw them as a danger, but also for profit. With the introduction of this project, the team in charge developed the role of “lion defender”, essentially offering money to local people to defend the animals and monitor the camera traps.

You can read this article from the University of Oxford to learn more about this project.

Oceanography and Marine Biology

Image by storyset on Freepik

When discussing climate change and biodiversity preservation, it’s essential to talk about the ocean and marine life.

As a major component of the global physical system, the ocean not only has a very significant effect on the climate but also all terrestrial and marine life. Despite this, there is so much still unknown about the ocean and the life within it.

The parameters assessed as we study the ocean can vary widely depending on the location and the time period. From the nutrients present in the water to the surface temperature, the data collected through field campaigns or autonomous platforms is still limited. Even satellites, which allow the constant collection of oceanographic data, are limited by the area they’re capable of monitoring at once.

Nonetheless, experts believe that artificial intelligence can be employed to analyze the existing data and, from there, fill the gaps of information and construct nonlinear models of the relevant ocean parameters.

This strategy could help us understand the changes occurring in the ocean today, as well as their impact on the climate and terrestrial and marine life. It could even be possible to use AI to predict future changes and weather events, which would be of particular importance to support marine activities like fishing.

NASA, for example, has a long-standing project for monitoring the ocean through its satellites. As mentioned on the project website, the data collected has applications across multiple areas, from the prediction of storms and hurricanes to tracking marine debris (as occurred when Malaysia Airlines Flight MH370 disappeared in 2014).

To find out more about the application of AI in this project, you can read this article published by the University of Washington.

Artificial intelligence can be applied in various ways to study the ocean and marine life, beyond modeling parameters that impact the climate or ocean currents. AI algorithms can be used, for instance, to monitor marine life or even assess the resources available in the sea.

As an example, we can look at a project created by Fisheries and Oceans Canada, Rainforest Connection, and Google, to help protect killer whales in the Salish Sea. While this was once a thriving species, today these animals have to face multiple threats to their survival, including a lack of prey, constant disruption by human activities, and water contaminants.

With this in mind, these organizations came together to develop an artificial intelligence system that helps to monitor and track killer whales which can help, for example, to efficiently find and treat injured whales. You can find more details about this project in this article.

If you want to learn more about the applications of AI in marine biology, we’d recommend starting with this article entitled “The Role of Artificial Intelligence Algorithms in Marine Scientific Research“.


Finally, the last field in life sciences that we want to discuss in this article is paleontology. At first, it may seem slightly odd to use artificial intelligence in this area; however, there are two main ways in which AI is proving to be a revolutionizing tool for paleontologists.

First of all, AI can be extremely helpful when studying fossils.

A crucial factor to consider during this process is the morphology of the fossilized organisms, given the array of information that can be obtained through it. For example, an animal’s morphology gives insights into characteristics such as its dietary habits or its gait.

Nowadays we have advanced imaging technology at our disposal to help reconstruct external and internal structures of fossilized organisms in 3D, without compromising these specimens. Nonetheless, it can be challenging to accurately differentiate between different materials in one sample if they have similar densities.

This is sometimes the case when studying fossilized bones that are encrusted in rocks, as the software may not be able to distinguish between them. Typically, in this case, paleontologists would have to manually separate the segments of the sample, which is not only time-consuming but also intensive.

As discussed in this Frontiers article, AI can simplify this process by conducting image segmentation faster and perhaps even more consistently, eliminating any potential bias from the researchers, who may interpret structures differently.

If you are interested in learning more about this application of AI, we’d recommend this research article, which described the use of a neural network for image segmentation.

Finally, as we discussed in our previous article “AI in STEM – Enhancing Scientific Research”, AI can create models from data that can be extrapolated into complex simulations. Within the field of paleontology, this principle can be applied to create simulations of wildlife and ecosystems from millions of years ago.

STEAM in AI Research and Build Program

At the DataEthics4All Foundation, we’re focused on helping the next generation enter and succeed in STEAM, Data, and AI fields. To this end, we started the STEAM in AI Movement, which includes a series of initiatives aimed at young people interested in these areas.

One of the main components of this movement is the STEAM in AI Research and Build Program. This is aimed specifically at high school students, offering them a chance to be mentored by PhD students and graduates working in STEM, while developing a unique project.

This opportunity allows students to learn more about the field they’re interested in, including potential career paths they could pursue, and receive important guidance to help them throughout their journey.

By conducting an original project, students also gain a realistic experience of what working in their area of interest is like so that they can make an informed decision about their future.

Our program could also be a starting point for students looking to enter STEM careers that wish to explore AI. Particularly, when considering the development of AI and its application across all STEM areas, learning more about it and gaining some experience can be extremely valuable for future professionals.

To learn more, visit our website or book a consultation with us.

Stay tuned for next week’s article where we’ll continue to explore the application of AI in STEM, this time exploring its impact across physical sciences, such as Physics and Chemistry.

Content Writer

Content Writer Intern for DataEthics4All Foundation | Freelance Science Writer | Creator of The Unfiltered Scientist | BSc (Hons) Biology

Post a Comment