Generative AI with Google Lens to detect counterfeits in pharma

Counterfeit pharma jeopardise patient safety, infringe on intellectual property rights, and undermine the credibility of the pharma industry. Current methods of combating counterfeits are often time-consuming, expensive, and prone to human error.

The recent government announcement that 300 samples would have QR codes has some restrictions. QR codes can be rendered inoperable by scratching them off using a sharp object or marker. It is useless if the QR codes are broken. What use does the label serve, and who is responsible for making sure the QR code is still working when it changes commercial hands? What assurance is there that counterfeiters who can easily mimic a brand’s packaging, name, and logo won’t also be able to mimic its QR codes?

A more efficient and accurate solution is required to identify counterfeit drugs and protect public health.

Combining generative AI and image recognition is probably the answer. With the use of Google Lens and Generative Artificial Intelligence (AI) technology, a complex system for assessing photo metadata (Photo Meta Data) on pharma packaging and verifying the authenticity in real time.

I’d want to make a recommendation for pharma companies, government officials, and customers to combine the strength of AI with picture recognition to create a foolproof way to restrict the risks connected with counterfeit drugs.

*In 2018, Google Images introduced some new features to its image search results. Next to a selected photo, the creator of the image, the credit line, and a copyright notice are immediately displayed. This works by reading the corresponding IPTC photo metadata fields embedded in the image file. On August 31, 2020, this feature was enhanced to also display a licensable badge above an image and a link to the licensing information.

System requirements

a. Generative AI: Implement advanced Generative AI algorithms that analyse visual patterns, text, and packaging features to identify discrepancies between genuine and counterfeit pharma packs.

b. Google lens: Utilise the powerful visual recognition capabilities of Google Lens to capture and interpret codes on the packaging swiftly.

c. Cloud-based database: Establish a cloud-based database containing comprehensive information about authorised pharma products, including packaging details, manufacturing locations, and authorised distributors.

d. Machine learning: Train the system using machine learning techniques to improve accuracy and adaptability over time.

System workflow

1. Generative AI will be used to create unique invisible image tags for each product (like UID). This can be done by training the AI on a dataset of existing batch numbers, allowing it to gener