By Graham McGibbon, Director of Partnerships, ACD/Labs
Many people’s perception of what a lab in the future may look like, or how it might function, comes from TV or movies where data is generated quickly and discoveries are produced in sudden “a-ha” moments. While in reality scientists that work in laboratories today, or have done so in the recent past, use experimentation processes that cannot match this expected speed. However, the goal of future technologies is surely to match this ideal.
So, what will a lab look like in, let’s say, 2030?
As one can imagine, no two days will be exactly the same—there will always be different techniques, different data, and different procedures in place. However, with Gartner recognizing democratization as one of the top 10 technology trends for 2020, we believe that artificial intelligence (AI), digitization, and structure characterization will come to the forefront of the laboratory, making the greatest impact on a scientist’s routine and the discoveries they make.
One can foresee scientists using AI to guide experimental design, which will improve efficiency in the laboratory and draw meaningful insights from myriads of data. Additionally, we think scientists will accelerate innovation in R&D through digitization efforts that streamline workflows and demonstrate data integrity. Finally, we believe scientists will reduce the risk and expense associated with research and develop new chemicals ideally with faster, cheaper, and more reliable structure characterization.
So, while their general mission will remain largely unchanged, how will these technologies and shifting workflows effect a scientist’s day-to-day routine? I took a guess at predicting what their workday may look like:
A future scientist’s day will start out like any other—after awakening and the morning routine, since the lab of the future is collaborative, they will scan not only world news that interests them but also any new scientific information denoted as very important in their team data sources. Maybe while they’re brewing or sipping a fresh cup of coffee they’ll be thinking about the first decision for the day’s plans: Are there any adjustments about the products we want and how we would make them? Options will come from a combination of proposals from AI systems combing through data overnight and thoughts emerging from the scientist’s own experienced and intuitive mind. Until further in the future, there will be some effort in the morning to merge those sources into the day’s experimental design.
For the foreseeable future some of the scientist’s work will often take place in a laboratory they physically visit. In the laboratory setting, AI will have already helped plan equipment and instrument availability along with the scientist’s schedule and be tracking the team members’ progress. All those involved in the experiment will be able to see details and updates on the day’s work. Meetings will already be scheduled and any pressing or urgent items in need of attention for human decisions will be flagged.
Since the lab of the future is automated and parallel, for producing new materials there will be processes set up, conducted, and monitored on equipment that are part of a wider inventory of available or potential raw materials and equipment. In combination with AI, the scientist will ensure the selection and set up of the preferred experimental process.
Setup requirements will vary based on the experiment, but the lab of the future is digital so we know one thing for certain—every experiment will use digital tools to control and track processes, recording human observations and capture instrument data. Some of these tools will be in the form of a tablet or hand-held device designed with the scientist’s workflow in mind. Perhaps it’s voice-activated or leverages video to record observations, but nonetheless, it will be capable of real-time tracking to make connections as data collection occurs. This will not only reduce the time it takes to document what was actually done and observed experimentally but it will be less subjective to help ensure repeatability.
In an ideal future, experiments would produce only the desired products in 100% yield but in the real world the outcomes will still depend on scale, duration, intricacy and uncontrolled variables in each experiment. So, the next decisions the scientist will face during the day are about any purification operations and detailed characterizations of material qualities. I would expect that a future scientist would conduct predictive simulations, review composition of matter analyses, or consider additional data collection depending on the experiment at hand and the initial characterization results. Most of these processes will be at least partially automated and able to share results so the tools used for decision-making will be designed to support collaboration and interaction. Future scientists will work collaboratively with one other, whether in the same laboratory or between different countries. Barriers to getting work done will be minimized as the appropriate data from each workflow pipeline is transferred into decision-support systems within minutes and any insights added will be available in real-time among collaborators.
A future scientist may peruse scientific journal articles related to their work or interests and also scan a summary report providing an overview of key observations collected and a review of what was accomplished to prepare for a team discussion of the day’s results and the next day’s experimental design. Any surprise results that emerged should be easily and quickly noted in the decision-support software or connected tools and important decisions will be made about uncertain interpretations, possible alternatives and the next day’s experimental design plans.
No future scientist will need to review his or her own notes for accuracy – those notes will be in a digital format and collected as the experiment is completed. This process will enable cost savings for the laboratory, speed up development, and minimize risk as any data discrepancies would be worked out in real-time. The technology storing this information will have cutting-edge security capabilities to support the integrity of the data and will only be accessible by the experiment’s collaborators. In an ideal future, robots will handle the end of day cleanup and begin setting up for the next day’s work. All maintenance would be scheduled and preventative. Future scientists will leave the laboratory knowing their day was productive and that the next day is already being planned out.
Thanks to AI, hours of human planning will be significantly reduced and the scientist can sleep soundly knowing things will be just as efficient the next day.
The scientist of the future will be able to handle data and make discoveries quicker than ever before thanks to the laboratory of the future. For an overview of what the laboratory of the future may look like, and how ACD/Labs’ capabilities and products may be helpful, read a previous blog post by Andrew Anderson.