Science Labs of the Future: Perspectives and Predictions

Laboratories are on the brink of a technological revolution. The implementation of Artificial Intelligence (AI) and automation into the laboratory environment will connect hardware, software and teams increasing productivity and ensuring an increased level of expediency and efficiency in the laboratory. It is expected that these technologies will shape the way research and development are conducted to the extent that the quality and yield of research will increase. 

Yet as machines and software begin to assume the roles that were previously in the domain of the scientist, the question of whether the role of the scientist will be diminished by these technological advancements needs to be addressed.

The introduction of technology into the laboratory environment isn’t a new phenomenon. Rather, it is quite common that new software and equipment is integrated into the laboratory to aid scientific research and development e.g. Illumina, ImageJ or Graphpad. 

What is different about the introduction of automation and AI into the laboratory environment is simply the scale of change that is expected to occur. Laboratories are structured, organised environments that are designed to ensure the precision and reproducibility of research. It is the nature of the laboratory environment which makes it the perfect candidate for automation and AI implementation.

The current system relies on scientists manually handling equipment and recording data. Large volumes of repetitive tasks are still the responsibility of the scientist. This can create an enormous strain on the scientist, who is not undertaking tasks that will further their research, but a large volume of menial, repetitive tasks that don’t require much thought or effort but take up a lot of time.

However, in the labs of the future, most tasks will be solved through the use of laboratory automation which includes the connection of instruments, devices and software. Information collected by the lab automation devices can then be processed and analysed by AI, which can make basic decisions in the laboratory to ensure experiments are run under the optimal conditions according to the results that have been analysed. 

So what will this mean for the role of the scientist? Will the occupation be replaced by the combination of AI and autonomous devices eventually?

MIT-IBM Watson AI lab recently released a report on the impact AI and surrounding disciplines will have on the nature of work. The research posits that “there is no question that AI and related technologies will affect all jobs”. Yet despite the emergence of these technologies and the subsequent long term changes they will bring to the nature of work, IBM research shows that few jobs will disappear entirely. Instead, a system will be put in place where technologies augment human ability not changing if we work, but how we work.

It is expected that the laboratory ecosystem will adapt to new technologies as the automotive industry adapted to the first introduction of robots in 1961. Tasks which do not maximise the time of the researcher will be transferred to autonomous collaborative robots (co-bots) which are designed to be fast, efficient and unintelligent. These co-bots can work alongside the scientist in the laboratory as they can sense the presence of a person and adjust speed and strength according to their environment.

Co-bots are also incredibly efficient at undertaking high volumes of repetitive tasks, they will work alongside scientists carrying out small manual labour tasks such as making small testing batches or retrieving certain items from an inventory. With the manual labour tasks transferred to the Co-bots, scientists will be able to dedicate their time to their more pressing benchwork duties. 

Robotics is not the only application of laboratory automation. In fact, there is a breadth of areas and equipment which are likely to become autonomous in the future, some that you wouldn’t even expect, like material containers. Autonomous devices will upload data generated from experiments directly onto a laboratories’ cloud or server, using a Laboratory Execution System.

Implementing this system ensures that the scientist does not have to manually record all research data in a lab notebook. Instead, the information will be streamed directly from the equipment onto a scientists’ Electronic Laboratory Notebook, (a digital software which provides a central platform for research data). Automation in the lab will simplify data acquisition and grant the scientist easier access to devices in the laboratory, which can be controlled and monitored remotely. This integrated system aids researchers who are trying to cure cancer, save the environment or simply discover more about the natural world. The introduction of this technology into the laboratory is something to be welcomed not feared. 

Put simply, automation will allow scientists to work smarter not harder.

In order to reach laboratory automation’s full potential, it needs to be combined with an AI or machine learning system. AI consists of statistical algorithms that learn with experience. These algorithms enhance the capabilities of automated systems enabling them to perform tasks that have traditionally required human cognition. It has progressed to the extent that now AI software can aid an experiment from inception through to its conclusion.

BenchSci, for example, provides a product that uses AI algorithms to assist in the selection of antibodies. Using the world’s antibody database the tool uses advanced machine learning to decode text and figures, reducing error and increasing the efficiency of collecting relevant data. 

During the experiment, AI and machine learning can derive meaningful information that goes beyond just recording the outcome of an experiment. Machine learning provides a systematic, less biased analysis that can be used to determine what laboratory procedures should be used in future experiments.

Through gathering and analysing relevant secondary data from the autonomous machines, AI can help maximise materials, laboratory items and researcher time. It uses continuous data from multiple monitors to determine how the experiment should be run, e.g. what environment or time will yield the best result. 

Once the experiment has taken place AI can analyse the data generated and produce new hypotheses from the experiment results. AI predictive abilities can also predict which materials will work best in the experiments, ensuring greater efficiency, precision and yields in the laboratory. In this way AI streamlines the entire experimental process, driving innovation and creating tremendous opportunities for those in research and development.

AI technologies are progressing at a phenomenal rate. Many complex tasks can now be carried out by AI algorithms, which minimise the risk of error and perform tasks quicker and more efficiently. AI is not something to be feared, it will not replace scientists. Rather, researchers need to harness the abilities AI offers to further scientific progression and alleviate the burden on scientists.

Laboratory automation and machine learning software will not diminish the role of the scientist. Instead, technologies will enhance it. AI and automation create effective methods for better AI-human interaction and collaboration. By eliminating the repetitive aspects of research, technologies will increase productivity levels, allowing scientists to direct their efforts to their research. With its growing prevalence, it is not surprising that AI and automation implementation has met some backlash, due to the breadth of sectors that it will affect, combined with the perception that these machines could eventually replace us.

However, trends suggest that in the research and development realm, the integration of AI and surrounding disciplines will augment human ability resulting in a greater number of global opportunities and an increase in scientific breakthroughs.