Artificial intelligence was a pretty universally feared and mistrusted concept until very recently, one that was largely ignored by the mainstream media, and relegated instead to the fringes of science fiction. In 2017, however, AI is everywhere. It's in our smartphones and our televisions, our cars and our vacuum cleaners, and even in our clothing. It's such a vast topic, in fact, that I honestly struggled where to begin this piece, as condensing the current trends and latent potential of AI in 2017 into under 2000 words would be nigh on impossible. So instead, I've chosen to step back, examine the wood from a distance, and pick out a few of the more interesting (relevant) trees to put a spotlight on.
AI in Advertising
Marketers might still be feeling out AI in terms of understanding the technology and the investment required, but many agencies are already developing their own dedicated AI divisions. Publicis Sapient, for example are one of the many agencies with their own AI division (this one headed up by Josh Sutton) mining the data they have on users through ad tech firms and social media platforms to create advertising faster and, theoretically at least, better.
It's not a technology limited to big agencies and big campaigns either. Erik Hallander, the Regional Mobile and Innovation Director at Isobar, has said that an increasing number of brands are testing the waters and finding that small and targeted outcomes can be surprisingly easy. He used the example of apparel brand North Face, which is using AI to recommend better products to its customers. Similarly, Facebook is drawing on deep learning to improve its automated system's analysis of users' behaviour, traffic and trends to recommend related content.
According to the 2017 Tech Toolkit from global marketing intelligence service Warc, 58% of global agency CMOs believe that, within the next five years, companies will need to compete in the AI space to succeed. So it's not just one lone voice shouting in the dark here, it's a good chunk of the industries leading experts who feel that AI is not only a good thing for marketers, but a necessary thing that everyone needs to get involved with sooner, rather than later.
Codec deliver AI driven content marketing to a dramatically expanded client list
Recently, Codec, the leading AI driven content marketing platform, announced 14 new customers including Unilever (Cornetto, Cif, Comfort), L'Oreal (Giorgio Armani, Shu Uemera, Vichy) and Universal Music. These brands are using Codec's unique content marketing platform to analyse people’s interactions with media over time to understand who their audiences are and what they care about when planning creative content. This understanding of audiences and what they want to consume effectively de-risks creative marketing initiatives in the planning stages. Content marketing is now a five billion pound industry in the UK alone and Codec’s AI platform allows marketers to create better (or at least more relevant) content by giving them an accurate prediction of what audiences care about before they plan and, rather than a static snapshot, it shows an understanding of people’s interactions with rich media over time. Since the company's launch in August 2016, Codec’s business grew by 53% and it is now working with a total of 21 clients across the UK, the US and Europe. If that isn't proof enough that AI, and the applications and platforms it enables, is growing and evolving with great haste, I honestly don't know what else would qualify!
AI in the Workplace
A recent piece of research from the online job board Jobsite, revealed that, despite media reports of AI impacting jobs, the UK is actually optimistic about the potential of artificial intelligence in the workplace. The survey of over 4000 Jobsite users showed surprisingly overwhelmingly positive attitudes to automation, with 54% believing advances will actually enhance their existing jobs, compared with only 33% who fear jobs to be at risk. This is completely at odds with the naysayers who feel the 'rise of the machines' could spell the eventual end of skilled, human labour.
Respondents believed that AI could enhance their work by helping them perform tasks quicker (63%) and allowing them to focus on more meaningful work (55%). Of the tasks expected to be assisted by automation, customer billing (52%), cybersecurity (49%) and administration (46%) came out on top. Whilst some surveyed showed concern for negative implications such as their skills becoming obsolete (37%), two thirds of respondents were already taking practical steps to pursue education in this area to improve their prospects.
Industries expected to be most affected by automation are manufacturing and banking. Meanwhile in sectors requiring a more hands-on, human approach such as social care and management, the risk appears to be comparatively low. The takeaway here? AI, at least as it stands currently, is being developed to help us, not replace us. Though it should be noted that we are still in the very early days here.
IBM Watson and the potential benevolence of AI
Easily the most famous and powerful AI in the world; the IBM Watson AI is being used by every forward-thinking industry with wallets deep enough to afford it, particularly adland. ADYOULIKE is the latest major marketing company to announce its partnership with the ultra-powerful system. The company, which is Europe’s largest artificial-intelligence-driven native advertising platform, marked its official US launch with the debut of the first server-to-server header bidding solution to incorporate Watson's AI-powered semantic targeting to assure that native ads are placed with the maximum contextual relevance. On a more grass-roots level, the Sears department store in the US is also currently piloting a program called the “Digital Tire Journey,” which allows customers to take advantage of IBM Watson to recommend which tires a customer should purchase. Not particularly exciting, maybe, but indicative of the many different levels and applications that Watson is being used at and for.
Perhaps more exciting, and important, is the fact that some of the biggest companies in medicine are using Watson to help solve medicine's toughest problems. IBM Watson for Drug Discovery and Pfizer are collaborating in the area of immunotherapy, which uses the body's own immune system to help fight cancer. Watson ingested 25 million Medline abstracts, over one million medical journal articles, data from 4 million patients, and every drug patent since 1861. The companies believe that the ability to recognise previously hidden patterns in the data will provide the next generation of targeted drug uses, while also discovering new uses for existing medications. This partnership was the result of a pilot program in which research information was fed into Watson, and it recommended a treatment that had already been suggested by Pfizer's own researchers. That potential treatment is currently in trials.
Medtronic is also working with Watson to develop a new generation of personalised diabetes management solutions. Using a combination of electronic medical records, health insurance records, and population data, the companies were looking to develop real-time personalised care. Using 10,000 anonymous patient records, the companies successfully developed the first-ever cognitive app, Sugar.IQ, that acts as a personal assistant and can detect patterns and predict diabetic events three to four hours before they happen, with a 75% to 86% accuracy rate. This helps patients better understand how their behaviour affects their glucose levels in real time. Finally, IBM and Quest Diagnostics developed a service that combined Watson's cognitive computing with genomic tumour sequencing and oncology diagnostics to launch IBM Watson Genomics from Quest Diagnostics. This service will provide specific recommendations and targeted therapies for cancer patients. So, if anyone ever tells you that AI is nifty, but isn't doing anything to further the betterment of mankind, point them in this direction!
Amazon Alexa and the 'mainstreaming' of AI
Amazon Alexa, the cloud-based AI voice service, is the first piece of technology in the last decade I can honestly remember my mum getting excited about and actively using on a daily basis without complaint. Considering we're talking about a woman who still refuses to text because she finds it too “impersonal,” that's surely something of a win? In order to take advantage of this immediacy and “personality,” Rant & Rave, the customer engagement specialists who work with half of the FTSE including Barclays, Sky and easyJet, has developed a customer feedback integration for the platform. For the brands using it, this will transform the way customer feedback is captured, enabling consumers to tell brands what they think of a product or service in real-time. Considering the existing size and variety of the install base of the platform, it's a solid concept that should gather some very useable results. The idea is that Rant & Rave, within Amazon Alexa, will sit in a brand’s 'skill' on the device, allowing customers to give real-time feedback about their experience. The integration means customers will be able to share feedback quickly and easily in their own words via their Amazon Alexa-enabled devices, wherever it is. For brands, this means customers voicing their feedback to Amazon Alexa, as they would to a friend or family member, providing valuable insights. It's these kind of simple, quick and infinitely useable features that Alexa was made for, and is the reason why it could act as the perfect gateway into AI for many users.
The Google AI almost as fast as the human brain
Deep learning machines (deep learning referring to a field of machine learning dedicated to replicating the thought processes of the human brain) have been generating incredible amounts of media buzz in recent months. Their abilities can allow them to play video games, recognise faces, and, most importantly, learn. However, these systems learn, on average, 10 times slower than humans. Now, however, Google has reportedly developed an AI that is capable of learning almost as quickly as a human being. Claims of this advancement in speed come from Google’s DeepMind unit in London, which says that, not only can their machine assimilate and act on new experiences more quickly than previous AI models, but it will soon reach human-level speeds.
Deep learning uses layers of neural networks to locate trends or patterns in data. If one layer identifies a pattern, that information will be sent to the next layer. This process continues until all the information is collected. The method mimics the processes of learning that occur in human and animal brains. Whilst this might seem like a science fiction nightmare on its surface, the possible benefits are enormous. Applications for advanced AI technology range from health and medicine to agriculture and even scientific research. As AI gets better at learning, it can be taught more ways to improve our lives, and that can only be a good thing.
Dyson and the AI vacuum cleaner
Last, but certainly not least, engineers at Dyson have hinted that the company could be moving into incorporating AI into its products in the future. In an interview with the Financial Times, the company confirmed it will be looking at how AI can be used in its 360 Eye robot vacuum cleaner, a circular machine which is able to clean floors on its own and which was launched last year. Dyson is looking into the more sophisticated use of vision robotics in its vacuum cleaners, which enables robots to move more accurately and respond to moving objects using camera sensors. What times we live in.
Benjamin Hiorns is a freelance writer and struggling musician from Kidderminster in the UK, who is just about spent on the subject of AI for the foreseeable future. If you want to hear from a few more qualified sources about their own thoughts on the immediate future of artificial intelligence, however, check out my companion pieces: Creative Opinions on AI. and AI Interviews: The Good, The Bad & The Creative. You can also check out our recent insight into How AI is Changing the Fashion Industry.