AI is not going to replace scientists, at least not any time soon. If anything, the more we learn about the actual abilities of today’s AI models, the more cautious we may be about trusting them. But at the same time, these models are clearly very powerful. It’s just a matter of figuring out how to best use them. That’s why Sapio leveraged its two decades of experience with AI-driven informatics to deliberately build ELaiN, an AI assistant established around the capabilities where AI reliably complements scientists.

Today’s cutting-edge AI, powered by Large Language Models (LLMs), can be used for a wide range of applications, with varying results. But what most of these applications boil down to is retrieving information and translating it into a new form.

If you ask an AI chatbot a question, and it can find the answer to a similar question on the internet, it will translate that answer to fit the new question. If you ask it to write you an essay on a topic, it will translate what it can find on the internet into the form of an essay.

The one thing that large language models can’t do is reason. That’s why the answer translated from a similar-sounding question is sometimes … a bit off. And it’s why the essays written by AI chatbots sometimes have glaring errors. Sure, these models occasionally appear to make connections and logical inductions. But it’s usually because they found the conclusion somewhere on the internet. More often, they just got lucky.

So, what good is an AI assistant who’s really good at translating but lacks basic logic? 

Well, it turns out that these skills are quite complementary to the skills of most pharma scientists.

If you ask a biologist what makes them a good scientist, their answer will probably include things like analytical and problem-solving skills, curiosity, attention to detail, and a strong foundation in scientific principles. They don’t need help with these things, and AI has none to offer, anyway.

On the other hand, many of the situations where biologists struggle, and would be happy to hand off the work, involve translation between scientific intuition and technical details:

Experiment Design and Management: Sapio’s ELaiN translates high-level descriptions of experiment designs into structured requirements, automation scripts, and inventory lists.

Querying Data: ELaiN translates scientists’ questions into detailed database queries to identify and retrieve exactly the data they need across multiple assays and experiments.

Data Integration: Once the data is found, ELaiN translates the original question into technical specs for how datasets should be combined and processed to extract the final answer.

Visualization and Analysis: Then, ELaiN translates follow-up queries into complex visualizations and analyses that allow scientists to dig deeper, limited only by their curiosity instead of by their technical skills.

This functionality is carefully designed to take advantage of what today’s AI is best at, while ensuring scientists remain the final authority, with full ability to review and verify ELaiN’s work. Plus, this is all seamlessly integrated into Sapio’s unified system, which combines an ELN and LIMS with a world-class data analysis platform.

With Sapio’s ELaiN, scientists can focus on what they do best while AI safely takes care of tasks that were previously tedious distractions.