Researchers Kha Tu Huynh and Huynh Phuong Thanh Nguyen, from the National University of Vit Nam Ho Chi Minh City have developed an artificial intelligence (AI) capable of detecting drunkenness by evaluating infrared images of human faces with an accuracy of 93%. Which, according to some analysts, unlike human police, would not be enough to arrest someone.
Kha Tu Huynh and Huynh Phuong Thanh Nguyen explain that early efforts to develop a way to detect drunkenness focused on eye condition, head position, or functional status indicators. However, these systems can be disturbed by other factors. The team points out that thermal imaging analysis offers a less ambiguous and non-invasive approach that could allow authorities to screen people in city centers or at events where alcohol is likely to be consumed. and where people can choose to drive home.
The team emphasizes the importance of any system designed to identify people who are intoxicated having a very low rate of false positives and false negatives. After all, a false negative could see an intoxicated person driving their car, while too many false positives would prevent sober drivers from using their vehicles and lead to public frustration and loss of confidence in the system.
There will always be a tradeoff in such a system, it would be better to err on the side of caution, but optimization of classification through larger training datasets on a diverse population of thermal images should to get closer to the ideal, which would of course be 100% accuracy, theoretically impossible to achieve, with zero false positives and zero false negatives.
A convolutional neural network can evaluate thermal infrared images of human faces and determine with 93% accuracy whether the person is drunk. However, there will always be a trade-off in such a system, so it is best to err on the side of caution, say the researchers. The use of algorithms for decision making has sometimes led to negative experiences. The Dutch scandal related to the use of algorithms is an illustration of this.
Indeed, in this case, it was said that the tax administration ruined thousands of lives with an algorithm. The Dutch authority responsible for personal data (Autoriteit Persoonsgegevens, AP) had imposed a fine of 3.7 million euros on the tax and customs authorities for the illegal processing of personal data in the fraud reporting system (FSV ) for years.
During its investigation of the FSV, the AP uncovered a long list of violations of privacy law, the General Data Protection Regulation (GDPR). It was a blacklist on which the tax and customs administration kept track of signals of fraud. This often had significant consequences for the people who were wrongly listed. As governments around the world turn to algorithms and AI to automate their systems, the Dutch scandal shows just how devastating automated systems can be, without the right safeguards.
Artificial intelligence to replace traditional methods?
Today, AI technology is used to perform tasks that range from the trivial to the monumental. From enhancing mobile photography through AI cameras to managing financial assets, AI has truly amazing applications in various areas of human life.
The omniscient, omnipresent, and near omnipotent power of AI has been questioned time and time again. The future possibilities of this technology can easily be compared to the myth of Pandora’s box. However, we know what was in Pandora’s box, the future of AI, on the other hand, remains quite unclear, which justifies the question we are faced with again and again, namely: will AI replace us human beings?
But as the applications of AI increase, more and more people are beginning to wonder if it will erase the importance of human skills and experiences in different tasks. What happens to a seasoned photographer’s skills and experience when the person holding an AI-enhanced camera just has to point and click? What happens to the skills and intuitions of an experienced stockbroker when AI is trained to predict stocks and help people make decisions based on that?
According to a report by the World Economic Forum, 85 million jobs will be replaced by AI-enabled machines by 2025. While this statistic may seem terrifying, the same report states that 97 million new jobs will be created by 2025 thanks to AI. So yes, AI will replace some jobs, but others are here to stay or will be created anew through artificial intelligence.
The assumption that AI will replace humans doesn’t sound so ridiculous. The positive aspects of AI are many, from speeding up mundane tasks to improving the internal workings of large industries. The analytical and predictive power of AI has proven to be a valuable asset to many industries. Whether predicting the weather or the stock market, the accuracy of the predictive power of AI technology is affecting many industries in positive ways. AI is not only used to analyze data and make predictions based on it, it is also used for more risky tasks like space exploration, sea exploration and many more.
Another positive point is undoubtedly the improvement in diagnostic and monitoring procedures in the medical industry. Doctors today are better equipped to diagnose a patient and monitor their condition, which leads to better and more accurate treatment. The effect of AI not only improves medical procedures, it also dramatically reduces the cost of medical treatment worldwide.
However, it should be noted that, since there is a positive side, there is necessarily a negative side to AI. From AI bias to social media manipulation, AI is affecting society in ways that were previously unimaginable.
The introduction of AI automation has brought about a serious change in the lifestyle of modern man. Tasks that once required manual labor can now be done without moving a finger. This means that users today have more free time and are looking for new experiences to fill this void. However, this search for new experiences is easier for some people than for others, which has a serious social and psychological impact on society.
One of the most egregious examples of the negative effects of AI is the incident involving a racist medical AI system that put millions of lives at risk. Racial biases in human society have entered AI through the datasets used to train this AI system. Millions of blacks thus suffered from their disease because they did not receive sufficient treatment.
In addition, there are much more sinister effects, such as manipulation of social media, which changes our perspective, lack of privacy and security, as shown by China’s social credit system, etc. The main problem with all these negative effects of AI is that we don’t even think about these effects until they are in front of our eyes. The negative effects of AI seem superficial when in reality it works on a deeper level, changing the fabric of society as we know it.
While it is true that the use of AI remains a debatable subject, it is no less true that the work of researchers Kha Tu Huynh and Huynh Phuong Thanh Nguyen, with AI capable of detecting drunkenness with 93% accuracy remains commendable. The system described in the International Journal of Intelligent Information and Database Systems could be implemented in places where drunk driving and intoxicated behavior are common problems. Every year, road accidents claim more than a million lives worldwide, many of which are the direct result of drunkenness.
Source: Vietnam National University researchers
In your opinion, will artificial intelligence replace human beings?
How do you assess the results presented?
Do you think 93% accuracy of an algorithm would be enough to arrest someone?
Is the hypothesis of seeing the police attributing a possible abuse of discretionary power to the account of an algorithm possible?
See as well :
The Dutch scandal is a warning for Europe on the risks linked to the use of algorithms, the tax administration has ruined thousands of lives with an algorithm
Meta’s algorithm has largely contributed to ethnic cleansing in Myanmar, according to Amnesty International, Rohingyas have been killed, tortured, raped by the thousands by security forces