Code

How Artificial Intelligence and Data Science Help Police and Medicine

How Artificial Intelligence and Data Science Help Police and Medicine

Free Python Course ➞ Mini-course for beginners and for Experienced coders. 4 cool projects in the portfolio, live communication with the speaker. Click and find out what you can learn in the course.

Learn more

When using the internet, we generate a huge amount of data. This applies not only to cookies, which many sites request permission for. Every action we take, including visiting web pages, interacting with content, and filling out forms, leaves a digital trace. This data can be used to analyze user behavior, improve the service, and personalize content. It is important to be aware of how exactly we share information and what precautions should be taken to protect your privacy online.

Social networks can reveal a lot about you, even if your profile is minimally filled out. There is a science that studies the relationship between a person's behavior on the internet and their psychological characteristics - digital psychometrics. It includes algorithms that analyze your posts, comments, likes, hashtags, and groups. This data can provide a deeper understanding of you than your immediate family. This is called a "digital footprint," and it can be used for a variety of purposes, including marketing and psychological research.

It's impossible to hide from technology, even if you decide to stop using your accounts. Social networks use tracking tools built into various websites. Recently, one of the most popular social networks introduced the "Off-Facebook Activity" feature, allowing users to view their activity history on third-party resources. This feature highlights the importance of being aware that your information can be collected and used even if you're inactive on the platform itself.

Are you considering giving up the internet completely to preserve your privacy? This solution won't bring the expected results. Even if you don't use online services, information about you can be accessed through relatives and friends who actively use social networks. A study conducted by German mathematicians has shown that a significant amount of data can be collected about a person even if they are not active online. Therefore, completely deleting yourself from the internet is no guarantee of protecting your privacy.

Research has shown that two people without social media accounts can be connected through a third person who does have an active page. Scientists claim that in 85% of cases, such connections can be established between unregistered users. This discovery underscores the importance of social connections and interactions in the modern digital world, demonstrating that even the absence of a social media account does not preclude the possibility of social contact.

Social networks know everything about us and even more

Scientists have long since begun to realize that data on user activities online can be analyzed and used. This topic goes far beyond targeted advertising. In 2009, American marketer Tony Fish published a book detailing the commercial potential of digital footprints. Analyzing this data opens up new business opportunities by enabling a more accurate understanding of consumer preferences and behavior.

The public began actively discussing digital psychometrics technologies after the 2016 US presidential election. Cambridge Analytica, which used digital footprint analysis methods to support one of the candidates, became the focus of attention. However, it remains unclear whether these technologies had a real impact on the election results.

There is an opinion that Cambridge Analytica used algorithms developed by Michal Kosinski without his consent. Kosinski is a recognized expert in the field of digital psychometrics, and his research is among the most cited in the world. This situation highlights the importance of intellectual property protection in science and technology, especially in the context of developing data processing and analysis methods.

Kosinski has devoted significant years to studying social networks, and the results of his work were published in a scientific article in 2013. In this study, he analyzes the influence of social networks on user behavior and their interactions in the digital space. Kosinski's work has become an important contribution to understanding the mechanisms that determine the dynamics of social platforms and their impact on society.

Kosinski was able to learn almost everything about a user by analyzing only a few dozen Facebook likes. His model identified characteristics such as race, sexual orientation, and political leanings of the account owner with a high probability of reaching 85-95%. This study highlights the power of algorithms for analyzing social media data and their ability to identify personal characteristics of users based on simple actions such as likes.

Predictions of religious affiliation, alcoholism and drug addiction, intelligence, and whether a person grew up in a single-parent family were less successful, but still successful. With a large number of likes (up to 300), the researcher could obtain more information about a person than their spouse or parents. This highlights the extent to which social media can reveal personal data and shape perceptions of a person, which is important to consider when analyzing user behavior and preferences.

To collect statistical information for his study, Michal Kosinski developed the MyPersonality project, which includes detailed psychological questionnaires. Users participating in the surveys often do not realize that they are granting the researcher access to data from their social media accounts by checking the appropriate boxes. This project has become an important tool for studying psychological characteristics and analyzing the impact of social media on participants' personal data.

The testing covered 7.5 million people, giving researcher Kosinski access to a vast database. This large sample allowed him and his team to develop highly effective algorithms capable of predicting hidden psychological and demographic personality traits. Using these algorithms opens new horizons in understanding human behavior and individual characteristics, which has important implications for various fields, including marketing, psychology, and sociology.

Kosinski is not the only researcher to successfully apply digital psychometrics. American scientist Tal Yarkoni analyzed the texts of 694 popular bloggers. Using the collected data, he developed a model capable of accurately predicting a person's personality traits based on just a few frequently used words. These studies highlight the growing interest in psychometric methods in the digital environment and their potential for deep analysis of behavior and personality traits.

Scientists are showing interest in the two billion images that users upload daily to public social media profiles. However, analyzing images is more difficult than text or likes. Spanish researchers have successfully applied deep learning technologies based on neural networks to image analysis. They developed a model capable of identifying basic personality traits based on a person's Instagram photos. This technique opens new horizons for studying the relationship between visual content and the psychological characteristics of users.

Cambridge Analytica: A Victim of Very Bad PR

Michal Kosinski and his team published the results of their research in the public domain, highlighting the potential risks of their algorithms being used by criminals. The researcher received offers from those wishing to use his developments for commercial purposes, but he declined, believing that scientific achievements should serve society, not generate profit. This approach emphasizes the importance of ethics in scientific research and the need for responsible use of technology.

In 2016, after Donald Trump was elected US President, the media actively discussed the theory of Cambridge Analytica's significant influence on the election. This British organization used digital psychometric methods to target election advertising on social media, which allowed it to effectively influence voters. Research shows that the use of such technologies has become an important element in modern political campaigns, changing the approach to interacting with the electorate.

This theory is largely based on a speech by Alexander Nix, the head of Cambridge Analytica. A video of his lecture is available for viewing on YouTube.

The company's operating scheme was based on collecting information about users through their actions on Facebook, including likes and subscriptions to communities. This approach allowed the creation of a vast database covering virtually all Americans. By analyzing this data, the company was able to identify user interests and preferences, significantly increasing the effectiveness of marketing strategies and refining the target audience for advertising campaigns.

Cambridge Analytica's algorithms analyzed the psychological characteristics of users based on available data. For this purpose, the OCEAN model was applied, which describes human personality through five key traits: openness to new experiences, conscientiousness, extroversion, agreeableness, and emotional stability. This model provides a deeper understanding of individual differences and human behavior, which can be used in various fields, including marketing and political campaigns.

Cambridge Analytica used detailed psychological profiles to create personalized advertising messages that appeared in users' Facebook news feeds. The emotional content of these messages was carefully tailored to the personality characteristics of each voter. This advertising strategy, known as microtargeting, allows for highly effective communication, increasing the likelihood of audience response and engagement.

Cambridge Analytica developed numerous ads targeting specific groups of voters. For example, users with patriarchal views were shown images of charismatic fathers. Those experiencing increased anxiety were shown ads depicting criminals breaking into homes. This approach to targeted advertising effectively influenced the audience's emotional state, increasing the chances of successfully attracting attention to political campaigns.

Alexander Nix claimed that the company's efforts increased Senator Ted Cruz's rating in the primaries from 5% to 35%. However, there is no convincing evidence that Cambridge Analytica actually had a significant impact on the election results. This influence may have been exaggerated by both the media and the company's management in order to attract attention and new clients.

Politicians have used similar technologies in the past, but they have not caused major scandals. The practical effectiveness of microtargeted advertising is likely only slightly higher than that of traditional television campaigning.

Cambridge Analytica was unable to perform the miracles attributed to it after Donald Trump's victory. The company's specialists also worked with candidate Ben Carson during the primaries, but his election campaign ended in failure. Despite claims about the influence of data analytics on election outcomes, real-world results show that not all strategies have been successful.

Following the Cambridge Analytica scandal, the company was shut down, and its management faced a lengthy justification process before the authorities and the public. However, the company's methods and the ideas of scientists can be applied to achieve positive results. It is important to consider how these approaches can be adapted for social and ethical purposes, such as increasing marketing effectiveness, improving customer engagement, and creating more targeted advertising. The use of data and analytical tools allows companies to better understand the needs of their audiences and offer them valuable solutions. Thus, Cambridge Analytica's experience can be used with an emphasis on ethics and responsibility, which will lead to positive changes in society and business.

But there are good intentions too

In the United States, social media scanning technologies are actively used to prevent crimes, including mass shootings and terrorist attacks. These measures are aimed at identifying potential threats and responding to them early. Analyzing data from social platforms allows law enforcement to track suspicious activity and prevent potential attacks, thereby enhancing public safety. Statistics show that virtually all potential terrorists post hateful manifestos, as well as messages and images featuring weapons, less than three days before committing a planned act of violence. This underscores the need for rapid social media analysis. This task is extremely difficult to accomplish without the use of artificial intelligence. Modern technologies allow for more effective threat detection and prevention, making the use of AI in online content monitoring not only desirable but vital. Tactical Institute specialists have developed a unique system that analyzes 500 million tweets daily for suspicious messages. If anomalies are detected, the system immediately transfers the data to the company's experts for thorough analysis, which takes only a few minutes. This technology enables the timely detection of potential threats and helps maintain safety on social networks.

Experts carefully analyze suspicious tweets, working around the clock to ensure safety. If a threat is detected, they promptly report the information to law enforcement. Twitter monitoring has helped prevent 14 mass shootings in American schools, highlighting the importance of such measures for protecting the lives and health of students.

Programmers at the Tactical Institute use IBM's Watson Analytics to search for messages. This service provides access to the powerful computing resources of the IBM Watson supercomputer and its artificial intelligence. The supercomputer is capable of finding answers to a wide range of questions, making it an indispensable tool for analyzing data and obtaining valuable information. Watson Analytics optimizes the search and analysis process, allowing users to effectively process large volumes of data and extract insights from it.

Analysis of social media content is used to identify people suffering from depression. Researchers have learned to identify such users by photographs, even if they do not show obvious signs of a bad mood. It was found that the profile images of depressed users are dominated by gray tones, and the photos are of poor quality and blurry. These people are often captured alone, and their facial expressions fail to convey emotion. This approach to content analysis allows us to better understand people's condition and develop effective support methods.

Depression can lead to serious consequences, including suicidal thoughts. However, these problems can be prevented if a person seeks help from a psychologist in a timely manner. In this regard, Facebook is testing a new system that analyzes messages and publications in groups. If the system detects phrases indicating suicidal thoughts, community moderators will receive an immediate notification. This will allow them to send a supportive message or suggest contacting a psychologist, which can be an important step towards recovery.

Analysis of social media messages can predict serious medical problems such as heart attacks and diabetes. Research shows that data collected from user posts and comments can serve as indicators of health status. Understanding emotional background and symptom mentions allows specialists to identify risks and suggest preventative measures. Thus, the use of social media in medicine opens up new opportunities for early disease diagnosis and prevention.

Data science specialists, together with medical professionals, conducted a study aimed at analyzing Facebook posts of heart disease patients. The study trained a system using data from the last five years before the onset of the disease. This helps identify patterns and risk factors associated with cardiovascular disease, as well as improve early diagnosis and prevention methods.

An algorithm identified approximately 200 marker words indicating a high risk of developing diseases. These words can serve as important indicators of health status and upcoming medical problems. Using such markers allows for the effective monitoring and analysis of patient health, facilitating early disease detection and timely intervention. Understanding the significance of these marker words can improve diagnostics and enhance the quality of medical care.

Vocabulary analysis can provide important insights into social tensions, economic difficulties, shortcomings in the education system, and crime rates. Algorithms exist that monitor the state of society in real time based on social media posts. These tools can identify key trends and issues, making them useful for researchers and government organizations seeking to improve public well-being.

Scientists have developed a "well-being map" of the United States, which illustrates the distribution of key health and socioeconomic indicators across different states. This map presents data on depression, unemployment, cardiovascular disease, diabetes, and factors related to economic and social development. This research helps understand how different regions of the country cope with health and well-being challenges and can serve as a basis for developing strategies to improve the quality of life of the population.

What about Russia?

We have similar projects that may interest you.

VKontakte has presented a new service that uses a neural network to recognize objects in photographs. This technology is capable of identifying thousands of different objects and warning users about potentially dangerous or illegal content. The new tool is aimed at increasing user safety and improving the quality of content on the social network.

The VKontakte system automatically classifies all photos uploaded by users into three categories: "safe," "moderately dangerous," and "dangerous." Content considered dangerous is subject to automatic blocking. However, the final decision to remove or block an account is made by a team of human moderators. If an account consistently posts content that violates the platform's rules, it may be suspended. This approach helps maintain the safety and comfort of all VK users by preventing the spread of inappropriate content.

In Russia, support for people with depression is being actively pursued. The social network VKontakte uses algorithms to analyze users' photos and hashtags. If the system detects signs of suicidal thoughts, the user will be offered assistance from professional psychologists from the "Your Territory" foundation. This initiative aims to provide timely assistance and support the mental health of users.

In Russia, there is a crime-fighting system called "St. George the Victorious." This system, similar to foreign solutions, uses artificial intelligence technologies for automated information analysis. It collects data from popular social networks such as VKontakte, Instagram, Facebook, Odnoklassniki, and Telegram. The use of AI allows for more effective detection of potential threats and crime prevention, making the system an important tool in ensuring public safety.

When calls for extremism, drug propaganda, or suicide are detected, the "George the Victorious" system immediately notifies the company's specialists. If an expert confirms the suspicions raised by the artificial intelligence, information about the distributor of prohibited content is passed on to law enforcement agencies. This effectively combats the spread of dangerous materials and protects society from potential threats.

The "SEUS" program, the predecessor of the current version, operates effectively in 20 regions of Russia and is actively used by law enforcement agencies. Its capabilities have helped prevent a number of crimes, underscoring the importance of this search engine in ensuring security.

The creators of the "George the Victorious" system emphasize that the platform works exclusively with publicly available information. This includes posts, comments, entries in open groups, hashtags, and likes. Private messages are not analyzed, confirming user privacy. It is important to note that open data analysis is sufficient for the system to function effectively.

Read also:

  • Artificial Intelligence, Machine Learning, and Deep Learning: What's the Difference?
  • Data: New Oil or Just Hype?
  • How Unit Testing Works in Python