Looking for life in the universe, however difficult, has been an interesting scientific problem for decades. Determining what is indicative of life (what to look for) has been highly contested. Here, we built an Artificial Intelligence model to detect if a sample is biotic or abiotic (living or nonliving) by creating a map of 280 diverse samples based on patterns in their chemical composition. Instead of looking for something specific within a sample, the model looks at each sample as a whole. The model can predict, with 87% accuracy, whether a sample is biotic or abiotic, and with 84% accuracy whether a sample is ancient biotic (coal/wood), contemporary biotic (carrots, etc.), synthetic (plastic, etc.), or from a meteorite. We also create a way to translate or adjust samples to match other samples taken from different machines to align and provide better results for some types of Artificial Intelligence models. Understanding how to shift the samples to align with the data already collected, we can apply data from other machines all over the world to improve the accuracy of current models, and we have built a model that can accept universally collected data regardless of any shifts.