AI tool can now easily identify synthetic drugs
Synthetic drugs are easier to detect with the AI tool
Researchers at the University of British Columbia have developed an artificial intelligence (AI) tool to analyze synthetic drugs. Until now, psychoactive “designer drugs” have been difficult to identify with standard drug tests. This new AI tool not only easily detects synthetic substances, but can also predict new chemical structures.
This AI advancement, called DarkNPS, could help develop more accurate drug tests with faster results. While current tests can take weeks or even months, this tool could provide results in just hours.
What are synthetic drugs?
Synthetic drugs are made in a laboratory to mimic the effects of other drugs that may be illegal or hard to obtain. These substances are created by changing the chemical structure of the original drug. In doing so, synthetic drugs circumvent existing drug laws and safety standards. Although often marketed as harmless “legal alternatives”, the lack of regulation makes the use of these substances extremely dangerous.
Currently, the 2 most common types of synthetic drugs in the United States are cannabinoids and cathinones. Synthetic cannabinoids, like K2 and spices, mimic the effects of THC which is the main psychoactive compound in marijuana. Manufactured cathinones, like bath salts, mimic stimulants such as cocaine, methamphetamines, and LSD. The effects of opioids, benzodiazepines and phenethylamines have also been reproduced. Use of these popular substances can cause negative health effects such as anxiety, high blood pressure, hallucinations, seizures, chest pain, extreme paranoia, and violent behavior.
The first US synthetic drug report came in 2008 when a shipment of spices was seized in Dayton, Ohio. To further avoid safety standards, synthetic substances are usually smuggled into the United States after being manufactured in another country. The packaging usually says “not for human consumption” or is intentionally mislabelled.
AI detection tool
Michael Skinnider, a medical/doctoral student at the University of British Columbia, worked on the development of the new AI tool for the detection of synthetic drugs. To achieve this, Skinnider and his colleagues collected more than 1,700 known synthetic drugs from around the world. Tandem mass spectrometry, or Tandem MS, has been used to examine the chemical structure of these substances. This technique breaks down the ions of the substance into smaller fragments. By doing so, the AI tool was able to find patterns between the chemical structures of these collected drugs. With this information, the AI tool was able to predict 8.9 million potential synthetic drugs.
After training this tool, the team of UBC researchers tested the AI technology with 194 new synthetic drugs. The DarkNPS tool was able to recognize 174 of the tested substances by looking for the most common chemical structure patterns. This 89% accuracy rate supports the idea that DarkNPS can be used to anticipate and identify future designer drugs.
Previously, the possibilities for new synthetic drugs were seemingly endless with an unknown amount of different chemical structures. With the ability to predict future structures, this tool could be useful in narrowing the search for harmful synthetic drugs. Additionally, DarkNPS could shorten the duration of investigations.
The impact of synthetic drugs
In addition to circumventing drug laws, synthetic drugs are cheaper to manufacture than other substances. Because they are chemically created, the supply chain needed to manufacture and distribute synthetic drugs is reduced, which can generate more profits. The potency of synthetic substances also contributes to their appeal to manufacturers who can sell smaller quantities with the same effects. These substances are also frequently mixed with other more expensive drugs to reduce production costs.
The mixture of synthetic substances is one of the main reasons why their use is so dangerous. This is especially true for fentanyl which is a synthetic opioid and one of the leading causes of overdose deaths in the United States. Originally created to treat cancer-related pain, fentanyl is 80 to 100 times more potent than morphine. This synthetic opioid is frequently used to cut heroin, unbeknownst to users. Because fentanyl is almost impossible to detect, the amount present is very unpredictable.
When individuals unknowingly consume a synthetic substance, there is a high risk of addiction, health problems and overdose. Just last week, it was reported that overdose deaths in the United States had reached more than 100,000 in the past year. This record number also indicated an increase in synthetic opioid overdose deaths.
The AI tool created by UBC could be useful in preventing future synthetic drug overdose deaths. According to the authors of the study, several authorities around the world have expressed interest in using DarkNPS.