Whenever you suffer from an infection, doctors generally prescribe antibiotics. The problem with antibiotics is that it’s a seemingly never-ending game of whack-a-mole, where what might work today, might not work the next. This is because bacteria mutates and evolves where it eventually gets to the point that the antibiotic you took might not work anymore.
Like we said, this can be a problem as it means that researchers have to constantly try to figure out their next step. However, it seems that the medical industry might have just gotten a helping hand thanks to AI, which has helped them discover potentially new powerful antibiotics that can fight against some of the more drug-resistant strains of bacteria out there today.
To train the AI, researchers taught it by showing it what kind of molecules kill bacteria. They tend fed the AI with a library of around 2,500 drugs and natural compounds, and how well these drugs work against the growth of the E coli bug. Once the AI was trained, the researchers then set it on a path of looking at over 6,000 compounds that are currently under investigation for treating various human diseases, and then getting to look for compounds that might seem effective.
They also taught it to think outside the box, where these compounds not only look effective, but are also unlike existing antibiotics. This in turn would allow the antibiotics to be more effective as it would allow it to combat bacteria that has yet to get a chance to develop resistance to. This seems to be particularly effective because according to the researchers, the process took a matter of hours.
The researchers are now hoping that they can apply the algorithm to find antibiotics that are more selective in the bacteria that they kill, and that eventually, they are hoping that this could also allow them to create brand new antibiotics from scratch.