SINGAPORE - Instead of having trained workers examine the glass bottles used for Japanese skincare brand SK-II's products, an automated system powered by artificial intelligence (AI) may be handling the job in the coming years, its parent company Procter & Gamble (P&G) told The Straits Times in a recent interview.
This came out of a collaboration between the American consumer goods giant and Singapore's Advanced Remanufacturing and Technology Centre (ARTC) in the new Jurong Innovation District, which is set to be a hub for advanced manufacturing in Singapore.
The centre is a unit under the Agency for Science, Technology and Research (A*Star).
The use of AI is expected to triple the speed of inspections on the packing line.
When adopted, it will be the first time P&G will use AI inspections on glass bottles, although it has deployed AI-based systems for inspections on other product lines such as diapers, said Mr K.C. Kasturi, P&G associate director for Asia beauty sector engineering.
P&G is among 70 members or so in the ARTC's membership consortium, which includes global multinational corporations and small- and medium-sized enterprises. Since joining, it has committed an initial investment of $2 million to drive projects like the use of AI in certain product lines.
"The new automated inspection system will allow us to raise the bar on quality by eliminating unnecessary human touches and potential errors," Mr Kasturi said.
Dr Cheng Fang, group manager of intelligent product verification at the ARTC, said multiple cameras are deployed along the AI-driven production line, with detection systems to pick up on scratches, dents and dirt on bottles simultaneously.
Mr Kasturi added that the new system provides faster, more consistent and more efficient inspection of empty bottles.
The technology is expected to be introduced progressively in the plant in Shiga Prefecture, Japan, where global SK-II production is run.
The system is slated to be shipped to Shiga by year-end for a trial and data collection, with a final installation slated for mid-2020.
Mr Kasturi said the current production line relies on staff to inspect empty bottles for defects such as scratches and dents, with the bottles sent to a packing line operating at 30 to 50 bottles per minute.
The automated system, however, will enable inspections of 150 bottles a minute, he added.
Asked what will happen to the staff who previously handled the bottle inspections, Mr Kasturi said: "The current resources will be reassigned to other meaningful roles when the new SK-II system is fully implemented.
"At this stage, this system is being sent to Shiga for further testing."