JAKARTA - Food security is a crucial aspect in the implementation of the Free Nutritious Meal (MBG) program. In the food distribution scheme for students, kitchen hygiene supervision, serving process, to delivery needs to be carried out strictly so that the risk of contamination can be suppressed.
Without a good control system, potential problems such as negligence in the use of personal protective equipment or the presence of pests in production areas can escape monitoring.
In this case, as a private partner of the MBG program, OVO - Grab Indonesia utilizes artificial intelligence (AI) and machine learning-based technology to monitor the progress of the MBG program in real time.
The MBG program run by this company is part of a social responsibility (CSR) initiative that is fully funded by the private sector.
Head of Safety & Quality Grab Indonesia, Sherylin, said that this system was designed to identify potential risks as early as possible, before the food was received by the students.
The use of technology was chosen because manual supervision was still considered to have limitations. With the support of a digital system, the monitoring process can take place faster and be data-based.
"We assess that manual monitoring has gaps. With technology, we can monitor in real time, and we can mitigate risks before the problem reaches the hands of students at school," said Sherylin when met at the MBG Command Center, Poins Mall, South Jakarta, recently.
In practice, supervision is carried out through CCTV cameras installed at three main points of the UMKM partner's kitchen, namely the cooking area, the serving area, and the handover point with the driver partner.
Sherylin explained that the device not only records activities, but has been equipped with AI technology that is able to recognize irregularities automatically.
"CCTV is not just recording, but AI technology has been added, so we can detect anomalies. We developed this technology ourselves," he said.
One of the system's capabilities is to detect compliance with the use of personal protective equipment (PPE) by kitchen staff, such as gloves, masks, head covers (hairnets), and aprons. If non-compliance is found, the system will send a notification to the monitoring officer.
"If the kitchen staff does not use it properly, we will immediately be given a warning later, such as there are several boxes, this is also one of the technologies we have created, so that it is easier for agents who monitor here, to see how the objects are detected and also accompanied by a confidence level, how confident is our detection of the object he detected," he explained.
Not only that, this technology can also recognize the presence of pests in the MBG kitchen such as rats, cockroaches, and lizards in the kitchen area. If detected, an automatic warning will be sent so that handling can be done immediately.
According to Sherylin, this system was developed internally and continued to be improved. He admitted that in the initial stage of the trial there were detection errors because the accuracy rate was still in the process of being improved.
"Detection of inaccuracies must have happened, especially at the beginning when this pilot project started. But we continue to make improvements every day by training more examples of what is wrong and what is right so that the machine learning is much smarter,"
"In addition, there are also our agents who are tasked with further reviewing to investigate people in the field, so there is still monitoring from the human," he concluded.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)
