There are several applications of AI in the pharmaceutical industry: optimization of production lines, lab expertiments, and safety.
Last week, Sanofi announced they plan to invest €60 million in new digital technologies by the end of 2021. Their goal is to adapt and modernize their entire global industrial network, including 75 plants around the world. The new technologies will allow to monitor the entire production chain in real time and perform predictive maintenance using connected sensors and artificial intelligence systems. Sanofi expects to increase its industrial competitiveness as well as the productivity of its employees.
On the other hand, Pierre Fabre organized an “AI Health Challenge” focusing on the prevention of skin cancer. Together with this initiative, in 2018 the group created a new European-wide Observatory of AI for healthcare professions. Its aim is to assess the penetration rate of AI among professionals and patients and draw an accurate mapping of the skills required.
These initiatives show that AI is becoming a major source of modernization and productivity for the pharmaceutical industry. A study conducted by Siemens Financial Services, presented at the World Industry Forum in Hanover, estimates a €60 billion potential gain from the digitization of production lines in the sector. In the same vein, in France, this transition would reduce costs between €1 and 2 billion.
Applications of visual recognition
Equipping production lines with sensors, connected objects, and computer systems would provide several benefits. First, it would increase productivity. Secondly, it would favor equipment planning and maintenance with automatic alerting devices, and continuous monitoring of the data.
These devices are mostly based on image recognition, a rapidly growing branch of AI. This technology offers many opportunities for the industry, particularly for the optimization of production lines, laboratory experiments, and safety.
Optimization of production lines
Visual recognition helps to set up predictive maintenance and quality control systems. In fact, cameras installed on the production lines can detect worn, damaged or defective parts on the products. For each anomaly, an alert is triggered, avoiding the current high downtime in pharmaceutical plants. A study by Siemens shows that digitization and data analysis could reduce these downtimes by 30-40%, thus significantly improving the overall efficiency of the equipment. Moreover, the law now requires a double human verification on the condition and cleaning of tools in factories. A visual recognition system is perfectly capable of doing this verification task, allowing factories to save time and gain in productivity.
Laboratory experiments
Concerning drug tests on animals in laboratories, the experimental periods are long and costly. By continuously filming animals’ behavior through smart cameras, it is possible to identify abnormal events in real time. Today, the observation is done manually. The new technology could save pharmaceutical companies time and allow staff to focus on higher value-added tasks.
Safety
Finally, visual recognition meets the security challenges of the pharmaceutical industry. Indeed, the handling of chemical, toxic products entails numerous risks of intoxication, injuries, and other damages. To protect themselves, operators must imperatively wear their personal protective equipment. However, this is not always the case. A visual recognition system can detect improperly worn or adjusted equipment, non-compliance with hygiene measures, and other potential risk factors such as a fire outbreak.
Conclusion
In short, the growing interest of major French pharmaceutical groups in AI and the wide range of applications that visual recognition offers to this industry are very promising. At Deepomatic, we recommend addressing the various issues that visual recognition can resolve in their entirety in order to optimize the solutions provided, create synergies between the various application areas, and finally, increase the productivity of companies.
Learn more about our solution here.
