The hashtag is dead, and machine learning killed it

In this series, retail experts at Shoptalk discuss the most pressing issues facing their industries today. Write your own #Shoptalk17 article here.

Our #selfies, our #tweets, and even our conversations are peppered with the ubiquitous pound sign. The symbol is so prevalent that “hashtag” was added to the Oxford English Dictionary three years ago.

But this piece isn’t about how popular the hashtag is - it’s about how we’re moving in a direction where it won’t even need to exist, thanks to smart technology called machine learning.

Computers today can “see” beyond the #hashtag.

The hashtag was created as a way to gather discussions and online exchanges around a certain topic so that online users all over the world could get involved in the same conversation. Hashtags used to be the only way to tie all the visual content on the web to the written content. But, recent advances in machine learning and image recognition technology have made it possible for consumers and businesses to be a part of the visual dialogue without relying on hashtags.

Through a subset of machine learning called “computer vision,” today’s most advanced visual recognition technology allows businesses to automatically understand visual content on a pixel-by-pixel basis, without the need for additional metadata, hashtags, or captions. In a time where customization is king, artificial intelligence (AI) can also be taught to understand specific brand logos (#nike), products (#applewatch), aesthetics (#iwokeuplikethis), and even the most abstract concepts (#lovetrumpshate).

Machine learning allows businesses and brands to make the most of their pixels.

The analysis and curation of authentic user-generated content (UGC) continue to be a primary focus for consumer brands in almost every industry. Brand managers want to know how consumers are engaging with their message. E-commerce operators want to include real world product shots on product pages to move users from browsing to buying. News organizations want to quickly find media being shared on the ground at a developing event. Hospitality marketers want to understand the most amazing moments of your trip. Social media managers want to find out what’s trending.

With the information gleaned from a set of pixels, businesses can power better media curation, data analytics, and product recommendations.

With visual recognition AI, businesses can build a more complete and accurate understanding of the consumer dialogue around their brand, message, and products. Hashtags can be a source of intelligence, but they are often noisy and messy. For example, if you search for #Oreo on Instagram, you don’t just see Oreo cookies - a large proportion of the results include pets named Oreo, a certain style of sneaker, and even sexually explicit images. Not exactly enlightening for a brand manager.

AI-powered visual recognition allows businesses to cut through the noise of poorly tagged UGC by looking at the pixels of an image or video. Businesses can really “see” what’s in UGC and improve their understanding of the conversation around their brand.

Brands can now better protect themselves and their customers from unwanted content.

Of course, in the age of social media, businesses aren’t just passively listening to consumer conversations on Twitter and cherry-picking a handful of pics to retweet. Today, more and more businesses are engaging consumers with contests, giveaways, social campaigns, and other marketing tactics that rely on users to actively submit and share content. This means that brands need a way to sift through all this inbound UGC in real-time and, sometimes, curate it instantly for public consumption.

AI-powered visual recognition can automatically moderate and filter harmful and unwanted content from UGC campaigns, protecting consumers and human curators from seeing offensive content.

Businesses who aren’t building machine learning into their media management workflows are already at risk of being left behind.

More than 2.5 trillion photos were shared in 2016, and the number is constantly on the rise. With so much new content to analyze and curate, businesses who aren’t building machine learning into their media management workflows are already at risk of being left behind.

While AI-powered visual recognition undoubtedly helps businesses make better decisions, increase conversions, and recommend better products, it can also be an expensive technology to invest in, particularly if your core business isn’t technology. Building machine learning technology in-house requires expensive upfront investments - research scientists, data scientists, special infrastructure, thousands of lines of code and data examples, not to mention continued maintenance and scaling.

Smart businesses today are already building their machine learning stacks with best-in-class AI technologies offered through plug-and-play APIs.

Luckily, there are other options. Smart businesses today have started to build their machine learning stacks with best-in-class AI technologies offered through plug-and-play APIs. Using an independent, third-party AI provider minimizes the upfront investment required while maximizing the utility of machine learning. It also abstracts away the technical minutiae of machine learning algorithms and gets rid of the infrastructure scaling costs associated with building and maintaining AI in-house.

The best out-of-the-box AI solutions will not only allow you to build AI into your business easily but also teach and train AI to learn new concepts and improve accuracy. After all, it’s not really “intelligent” technology unless it can make itself better.

Providing consumers an AI-first experience will soon become as necessary as providing a mobile-first experience.

As machine learning becomes involved in every aspect of our daily lives - from self-driving cars to self-paying checkout systems to self-regulating smart homes - consumers will grow to expect businesses to provide an AI-first experience. The hashtag may soon become a relic of the past, but if you start investing in AI now, your business won’t have to.

Clarifai is an artificial intelligence company that excels at visual recognition. Our products make it easy, quick, and inexpensive for businesses to innovate with machine learning, go to market faster, and build better user experiences. We make teaching AI just as accessible as we make using AI, which is why our technology is the most customizable, unbiased, accurate solution on the market. Find out more @Clarifai.