How much of an ally can AI be in the fight against counterfeiting?

By embracing the power of artificial intelligence, fashion brands have the opportunity to stay ahead of the counterfeiting landscape and protect their customers from fraud. 

The ever-evolving landscape of counterfeiting can be a challenge for fashion brands to stay ahead of the game, but AI-driven solutions offer them an invaluable advantage. By utilizing AI technology, fashion brands can protect their customers from the threat of fraud but also safeguard their products and maintain their competitive edge in the market. So, artificial intelligence is becoming increasingly important in the fight against counterfeiting – it’s no longer just a nice-to-have, but a must-have tool for any fashion business looking to stay ahead in this ever-changing industry. As Forbes says: “If you truly want to protect your brand, start with AI”.

How counterfeiting in fashion has evolved

According to the OECD, global losses from trade in counterfeit and pirated goods are estimated to be around $500 billion a year. In 2018, the fashion industry accounted for 34% of counterfeiting cases worldwide.

Counterfeiting has been around since the dawn of commerce, but the digital age has enabled counterfeiters to reach a much wider audience. With online platforms making it easier than ever to buy and sell fake items, fashion brands can quickly find their creations being replicated by fraudsters. Moreover, the skills of counterfeiters have improved over the years, so much so that today original and fake items appear very similar, certainly indistinguishable to the untrained eyes of a consumer. But even insiders can sometimes fall into error. 

However, the fight against counterfeiting is not only a question of brand protection or lost sales. Fake products are made of low quality and often contain dangerous substances. This poses a threat to the environment (due to the high number of items that quickly end up in landfills) and to the health of consumers.   

For these reasons, fashion brands need an agile approach that incorporates AI technology to detect any potential counterfeits before they make it into the market.

AI and counterfeiting: what brands can do

AI algorithms have become incredibly sophisticated in recent years, with the ability to rapidly process large amounts of data from different sources (such as product images, customer purchase data, etc.)

There are two main areas of application for artificial intelligence in the fight against counterfeiting: monitoring online markets and helping product controllers. This is possible thanks to the following AI functionalities. 

1. AI and counterfeiting: the role of image recognition

AI-driven image recognition is the basis of many counterfeiting detection solutions.

AI is able to find product images online that are similar to those in the company's databases and compare them to detect minor discrepancies (and intellectual property infringements) that could indicate a fake item. For example, AI-driven solutions can recognise patterns that may indicate a potential knockoff, such as mismatched logos or discrepancies in stitching quality.

AI algorithms can also detect any changes to product labelling—like incorrect fonts or colours—that might signal a fake product. This helps brands identify suspicious listings before they even reach consumers, allowing them to quickly remove infringing products and protect their customers’ trust. But AI algorithms can also monitor online markets and detect other suspicious activity, such as products that look original but are sold at a lower price.

2. Natural Language Processing to monitor online conversations

AI-driven solutions can also track patterns of behaviour in order to discover potential fraudsters before they cause damage to brands. By monitoring online marketplaces and social media platforms where counterfeiters typically operate, AI algorithms can alert fashion brands to any suspicious activity before it takes place. Indeed, by leveraging natural language processing (NLP) technologies, fashion brands are able to scan through large volumes of text from multiple sources, such as social media posts or online reviews, to identify words or phrases that may suggest counterfeit activity. This information can then be used to alert companies about potential violations and take action against offenders accordingly.

3. Trend analysis: the secrets behind behaviour

Machine learning algorithms are constantly evolving and improving their capabilities to detect counterfeits, even the most sophisticated and difficult to spot. By analysing data, AI-powered solutions are able to quickly identify anomalous trends that could indicate a fake item. For example, if a fashion brand notices that a large number of products with their logo are being sold on an online marketplace, they can use AI to investigate whether these are genuine items or not. If the AI detects that the majority of these products are fake, the brand can then take steps to remove them from the platform and protect their customers. 

4. The importance of prevention

Additionally, AI can also be used to help fashion brands prevent counterfeiting before it even happens. By analysing past instances and understanding the motives of counterfeiters, AI can provide brands with insights into how best to protect their products from being replicated. This can be done by tracking product designs, supply chains and production processes in order to prevent any potential risk.

5. Beyond the product: labelling and packaging

In addition to monitoring markets for potentially counterfeit items, AI can also be used to help fashion brands develop more secure product labelling and packaging. AI-driven solutions can be trained to distinguish authentic product labels from fakes which are indistinguishable from each other even to a trained eye making it much harder for counterfeiters to copy a brand’s products. 

6. Fighting counterfeiting along the supply chain

AI can also be used to monitor the entire supply chain and track products from production to distribution, and detect any unusual behaviour, so fashion brands can detect any illicit or non-conforming at any stage of the process.

7. Overcoming human limits and helping product controllers 

Of course, AI-based tools can help product controllers in their daily tasks, such as examining hundreds of items per day to determine whether they’re genuine or fake. As we have said, AI algorithms can analyse images and accurately identify any discrepancies that would have otherwise gone unnoticed by the human eye. Moreover, AI can also be applied to analyse textiles for signs of authenticity, such as detecting irregularities in embroidery or other features that suggest a garment is not genuine.

8. AI and counterfeiting: authentication solutions

Finally, artificial intelligence-based authentication solutions allow customers to verify the authenticity of a product before making a purchase, e.g. by using smartphone applications that scan barcodes or other labels on garments, as is the case with connected products by Certilogo. This way, buyers can be sure they’re buying genuine items.

How to choose an AI protection solution

As Mark Lee or Forbes’s article “How Artificial Intelligence Helps Protect Intellectual Property” explains, there are at least 3 elements fashion brands should consider when choosing an AI solution for brand protection:

• The type of AI: Not all artificial intelligence is the same. Some are only for image analysis, others for text analysis, and others do both. It is important to know this in order to choose the best solution for a company’s needs.

• The supplier: Does the supplier build the artificial intelligence solution in-house or does it rely on third parties? Is the solution customisable and modifiable over time or does it have a rigid structure? This affects both security and adaptability to change.

• Machine Learning: Does the AI solution have a machine learning system able to analyse usage data and automatically improve its ability to detect counterfeits? Or is human intervention needed? These elements affect costs and also future efficiency. 

Certilogo

Certilogo has been a pioneer in adopting artificial intelligence to enable consumers to verify the authenticity of products from numerous fashion brands through its unique tagging methodology.

On average, we process a new authentication request every 4 seconds.

How do we do this? By analysing the data acquired in real-time through a proprietary Fraud Detection System, which is based on Artificial Intelligence and is able to recognise fake products, as well as any non-compliant distribution of authentic products. 

The quality of our system is also certified by the fact that judges in the annual Awards AI  competition have chosen Certilogo as the winner in the category “Best Use of Artificial Intelligence in Fashion” for 2018-19.