The term “Artificial Intelligence” (AI) is loaded with a set of assumptions taken from decades of science-fiction movies about robots taking over the world. Far from being part of a fanciful story about your microwave developing feelings though, AI is already common fact of modern life. Although dreams of robots who can take our place at work while we spend our days golfing and going for long lunches may be some distance in the future, what computer scientists call “narrow” or “weak” AI is already integrated into our daily lives.
What Is Narrow AI?
AI has been a developing field of technology since the mid-1950’s, when Allen Newell, Herbert Simon and Cliff Shaw created “Logic Theorist”, a computer program which was designed to reproduce human problem-solving skills. Since then, these programs have become ever more complex, ranging from champion chess machines to the algorithms which run the stock market. Chances are that if you’ve navigated anywhere using a Sat Nav, flown in a plane big enough to rely on autopilot – even if you’ve simply been protected by spam filters from a phishing email, you’ve used AI.
These weak AI examples can be looked at as computer programs which are explicitly designed to receive information and then to interpret it and solve a particular kind of problem. In the case of your Sat Nav – you’ve input your route and are driving happily along with your estimated time of arrival. However, along the way there’s a crash ahead of you and although you cannot know about it, the information is relayed via satellite to your device, which then solves the problem of a waiting time by adapting and re-routing you. This is AI in action.
What Does This Have To Do With Brand Protection?
Increasingly, the problems posed by counterfeiting are online. The scale of the problem varies by brand, country and platform, but it’s clear that some brands and products are being disproportionately affected by counterfeit online sales in a way that they wouldn’t be by physical shops or sellers. Early in 2018, the US games company Asmodee estimated that for some of the games they sell, around 70% of US sales are counterfeit. Clearly not all brands are affected in such a dramatic way, but it indicates the seriousness of the online issue.
It’s here that AI can make a difference. There are essentially two problems and two AI-related solutions. For brands who are new to dealing with anti-counterfeiting the amount of counterfeit listings on the Internet can be overwhelming and altogether too much for the available human resources to continually scour and remove. In this instance, some AI technologies may be used to perform web-crawling searches which use keywords and metadata to return results. Once search results have been checked and assigned scores then the programs can be taught to utilise this information about which adverts are more likely to be genuine or not and to differentiate listings and return more accurate results. Fundamentally this would reduce the need for human time spent searching and thus be more cost effective.
Surprisingly, however, the need for AI solutions may actually be more pronounced for those brands who already have existing brand protection strategies and a more mature anti-counterfeiting program in place. For brands who have already spent time eliminating counterfeit product listings there will be an easily identifiable change in the way fakes are described. To take a specific example, for football clubs one thing that commonly happens is that sellers of fakes quickly realise that their listings are removed when they include a team name in the description of the product. As a result, the listings are reintroduced with a slightly different description using more generic terms such as “Spanish football team kits 2018/2019,” for example. This makes these listings harder to find for both human and machine searches, yet they still have a place online and may be found by consumers. In this instance one of the developing AI technologies for the future is likely to be image recognition. With image recognition it will be much easier for brands to seek out and remove those listings which use their IP to sell counterfeit products, thus combining all the available technologies in the most cost effective, time efficient way possible.
Whichever kind of brand you are, whatever kind of online brand protection problem you may have, why not speak to the team at Pointer and make your anti-counterfeiting work as cost effective and time efficient as possible?