We will never know if better chemistry would have made 2001: A Space Odyssey’s AI villain less paranoid, but scientists in Belgium have built their own artificial intelligence-powered bad trip warning system, dubbed HEMP9000.
A team of researchers led by Dr. Christophe Stove of Ghent University has developed a machine-learning test which accurately screens for potentially dangerous synthetic cannabinoids. The team’s study, published in the American Association for Clinical Chemistry’s (AACC) Clinical Chemistry journal, claims the computer test can detect synthetic cannabinoids with greater ease and lower cost than standard methods like mass spectrometry.
Detecting synthetics is of growing importance. Up to 30 new varieties of designer drugs meant to mimic the effects of cannabinoids appear every year, according to reports.
These synthetics can have harmful effects. To produce the effect of potency, many synthetic cannabinoids target the CB1 cannabinoid receptor in the brain. Natural cannabinoids are generally well tolerated by these receptors, while synthetic cannabinoids are more frequently associated with acute and chronic toxicity.
In the central nervous system, powerful synthetic cannabinoids can induce potentially dangerous psychomotor impairment, agitation or psychosis. Regulation of neurotransmitter release by CB1 receptors is likely responsible for these potentially life- and/or mental health-threatening conditions.
Detection could help health care providers, officials
By making it simpler to identify new varieties of synthetic cannabinoids, the machine learning-based test could be a boon to health professionals and lawmakers, Stove and his team claim. For example, hospital emergency room screenings using the new technology could pinpoint proper medication for patients affected by bad synthetics. Health officials could use HEMP9000 to more easily identify new synthetic cannabis formulations to corral the designer drug market.
The current method for identifying these drugs is to use liquid chromatography high-resolution mass spectrometry, which is time-consuming and expensive, the scientists said.
The new test employs a bio-assay that uses cells designed with a CB1 cannabinoid receptor on their surfaces. When these cells are exposed to patient samples that contain synthetic cannabinoids, the cells emit fluorescent light. The computer model reads the level of luminescence to determine the presence of synthetic cannabinoids.
Stove and team’s machine-learning model was able to find 141 of 149 samples identified as synthetic cannabinoid-positive through mass spectrometry. The resulting sensitivity of 94.6 matches the accuracy achieved when mass spectrometry results are tested against interpretation by a human expert.
“In conclusion, the bio-assay continued to demonstrate outstanding performance, confirming its potential as an ideal untargeted screening assay, capable of sensitively and universally detecting new circulating [synthetic cannabinoids],” Stove