Digital drugs and role of AI in combating them

Observer

13-02-2025

Two weeks ago, I met Omani researcher Dr Mohammed al Husseini at the Muscat Global Knowledge Dialogue hosted by the Sultanate of Oman. We discussed digital drugs and their unknown risks. In this context, Dr Al Husseini and a group of researchers published a paper with IEEE last year. Titled 'Detection of Digital Drugs Using Artificial Intelligence Deep Learning', the study examines how advanced deep learning models can detect 'digital drugs', auditory effects produced by delivering slightly different frequencies to each ear, which in turn alter mental status.
Given AI's ability to address such digital risks, the study confirmed the effectiveness of Inception MV4 in detecting digital drugs.
The researchers employed the Inception MV4 model to analyse and characterise audio data. Their dataset comprised 7,000 audio files: 5,000 containing digital drugs and 2,000 original files (including music, animal sounds and Quran recitations). The methodology converted the audio into visual representations (frequency spectra) using the Short-Time Fourier Transform (STFT), a mathematical technique for analysing audio and time signals. The Inception MV4 model was then trained on these images to classify them. The results revealed an impressive classification accuracy of about 99.97 per cent using three different learning rates, with Inception MV4 outperforming the ResNet-50 model in terms of accuracy and true positive rate.
Given AI's ability to address such digital risks, the study confirmed the effectiveness of Inception MV4 in detecting digital drugs. The researchers recommended expanding the dataset and exploring additional optimisation techniques to further enhance performance and practical applications. They also noted that this system could be used to monitor and regulate online audio content to identify harmful auditory effects.
The discussion above exposes a dangerous phenomenon of digital addiction that threatens human mental and cognitive well-being, while simultaneously highlighting AI's role in uncovering these hidden digital substances. This summary underscores the importance of employing AI to tackle this growing issue. In exploring digital drugs, we must consider psychological, social and scientific perspectives. I recall an article I published in the scientific supplement of Oman newspaper, where I discussed how certain frequencies shape 'cosmic energy', mostly with positive effects. However, it is essential to distinguish between positive and negative auditory frequencies. In that article, I noted that scientific studies confirm various effects, psychological, spiritual (consciousness) and physical, from specific auditory frequencies, such as 432 Hz.
In our current discussion, we assert that we live in a world of diverse energies and frequencies. Certain frequencies, when manipulated (for example, by altering the sounds heard in each ear), can produce what we call 'digital drugs'. Scientifically known as 'binaural beats', some claim these beats can stimulate the brain to induce effects similar to those of narcotic drugs. Proponents argue that listening to two different frequencies in each ear can lead to brainwave entrainment, resulting in relaxation, euphoria or other altered mental states. However, studies indicate that exposure to such digital sounds can foster an addiction akin to that caused by chemical psychoactive substances. Their risks extend to physical and psychological harm by implanting subconscious suggestions that drive individuals to behave contrary to their normal patterns and societal norms.
Digital drugs are not a new phenomenon; researchers have examined them for decades. Recently, their covert integration into certain Quran recitations has raised alarms in several countries, prompting significant efforts and expenditures to combat these harmful auditory materials and treat those affected.
In addition to public awareness about the rising risks, especially with headphones that can alter frequency levels in each ear or negative auditory frequencies that require no such devices, we also need advanced digital tools to track and detect these substances. The surge in scientific studies proposing AI models, including Dr Al Husseini's work, underscores this need. The solution to curbing the spread of digital drugs lies in deploying counter-digital systems based on AI models that have proven effective in detecting and classifying harmful auditory frequencies from those that are benign.