What is Artificial Intelligence (AI)?
In our previous webinar, Chatbots & Conversational Commerce, we introduced two early chatbots, ELIZA and Parry. ELIZA was developed by MIT professor, Joseph Weizenbaum, in the early 1960s to parody “the responses of a non-directional psychotherapist in an initial psychiatric interview.” Weizenbaum had, in fact, set out to prove that interactions between humans and computers would only ever work on a superficial level. He surprised himself when he found that many who interacted with ELIZA began to form a recognizable emotional bond. The next major chatbot evolution came in 1971 with Parry, a natural language program that was developed to simulate the responses a paranoid individual would give. Building upon ELIZA’s advances, Parry was able to pass the Turing Test - a test developed by Alan Turing in 1950 to test a machine’s ability to become indistinguishable from a human. Subjects interacting with Parry and actual human responders were unable to distinguish, with more than random accuracy, Parry’s responses from those of a paranoid human.
There is little debate over the impressive computational advances heralded by ELIZA and Parry. However, there has been plenty of discussion since over what actually constitutes artificial intelligence. Philosopher, John Rogers Searle, argued in his 1980 paper, Minds, Brains, and Programs2, that “a program cannot give a computer a “mind”, “understanding”, or “consciousness.” Though ELIZA and Parry’s responses were human-like, enough so that they could built empathy or even replicate a human’s response, ELIZA and Parry themselves did not understand their own responses. They were simply manipulating symbols to pass the Turing Test and fool humans. This brings us to the state of artificial intelligence today as we enter 2018 and look toward the next decade.
Today’s descendants of ELIZA and Parry are chatbots such as 1-800 Flowers’ which allows you to order flowers through Facebook Messenger or Taco Bell’s Tacobot which allows you to order tacos through Slack. This level of artificial intelligence is described as “Weak” or “Narrow AI.” Tacobot can perform one function; order tacos. It does not possess human-level intelligence nor does any form of AI as we know it today. Even Google’s AlphaGo which beat a human professional player at the notoriously difficult and abstract board game Go in March of 2016 and beat the world’s number one player earlier this year, is a form of “Narrow AI.” It can play Go better than any human, but that’s the only thing it can do. For now, humans posses a unique and special level of intelligence, but recent developments are beginning to cut into the uniqueness of our minds.
Let’s now discuss some of these developments or AI building blocks and what they mean for the next evolution of artificial intelligence.
Five Elements of Artificial Intelligence
Machine Learning (ML) is “a current application of AI based around the idea that we should really just be able to give machines access to data and let them learn for themselves.” The term was coined by Arthur Samuel at IBM in 1959 and is related to computational statistics and mathematical optimization. Though ML itself has become a buzzword, it is the key to unlocking insights from another buzzword, Big Data. The algorithms powering ML could potentially sift through all of our data to make predictions and deliver analysis that the human mind could not alone decipher.
Deep learning describes computational processes build to mimic the human mind’s neural networks and allows for nonlinear analysis. These neural networks take “metadata as an input and process the data through a number of layers of the non-linear transformation of the input data to compute the output.”
Natural Language Processing
Natural Language Processing (NLP) is “a way for computers to analyze, understand, and derive meaning from human language in a smart and useful way.” John Rogers Searle took issue with ELIZA and Parry’s ability to understand language. They were merely responding to the input of symbols with symbols that meant nothing to them. NLP would enable ELIZA and Parry to understand language in the same way a human can. Siri, Alexa, and Cortana use NLP to recognize speech and enable voice search.
Natural Language Generation
If NLP is the input, then Natural Language Generation (NLG) is the output. NLG enables programs like ELIZA and Parry to engage in an actual conversation. NLG “assess the data to identify what is important and interesting to a specific audience, then automatically transforms those insights into...insightful communications.” Digital assistants like Siri use NLG to respond to task requests with intelligible sentences.
If you have spent any time on social media over the past few years, visual recognition is already a concept you are familiar with. Social networks use visual recognition to analyze images for visual content. It is how photos that you post get auto-tagged. It is also how Snapchat’s lenses and filters work. Here’s IBM Watson’s demo of its visual recognition capabilities: IBM Watson Visual Recognition. It allows users to upload any image and tells you what Watson sees.
Machine learning and artificial intelligence have phenomenal potential to simplify, accelerate, and improve many aspects of our lives.
Tony Bradley, Forbes
The Future of AI
The five elements of AI are all being advanced by research teams at major organizations such as Google and Facebook. Further, as machine learning and AI advisory group, Topbots, highlights, the landscape of enterprise organizations pursuing artificial intelligence technologies is becoming an increasingly crowded and competitive space. Though we have yet to achieve artificial general intelligence (AGI) or human-level intelligence, as we discussed earlier, the field is advancing at an accelerated pace, recall again Google’s Ray Kurzweil's prediction that we will achieve AGI by 2029. This rapid pace is why it is worth hearing from those who are calling for caution as we embrace AI, namely Elon Musk.
It may come of a surprise that one of the loudest cautionary voices is one of our foremost futurists. Elon Musk, the founder, CEO and CRO of SpaceX and co-founder of Tesla Inc, is no stranger to the benefits of AI. Telsa’s autopilot features use AI and it has been reported that the company is developing its own AI processor for its self-driving cars. Further, Musk was an early investor in the AI company, DeepMind, which was acquired by Google in 2014. Though Musk claims he got involved more to keep tabs on how quickly AI was evolving. Which, he says, is happening at a pace much quicker than people realize. He cautions that many of his peers, like Larry Page at Google, are naive in their boundless pursuit of AI and may “produce something evil by accident” possibly even, “a fleet of artificial intelligence-enhanced robots capable of destroying mankind.”
To prevent the apocalyptic possibilities that artificial general intelligence may herald, Musk has started a nonprofit artificial intelligence research company, OpenAI, with the aim of promoting the safe and ethical use of artificial intelligence. Musk is not alone. Many of our most distinguished thinkers and technologists including Stephen Hawking and Bill Gates are urging caution.
As marketers, it is important to be aware of these cautions. We are early adopters of AI technologies and as we will soon discuss, marketing, as a profession, will be profoundly disrupted by AI. Our decisions on what technology to purchase or build for our organizations, will, without any sense of hyperbole, affect the future of humankind.
“Everything we love about civilization is a product of intelligence, so amplifying our human intelligence with artificial intelligence has the potential of helping civilization flourish like never before – as long as we manage to keep the technology beneficial.“
Use Cases & Examples of AI & Machine Learning in Marketing
As marketers our goal is always to reduce friction; on our websites as our audience searches for answers, with our content etc. In one sense, the singularity, the merging of human and machine intelligence is already here. We already interact with our smartphones as extensions of ourselves. We have to remember less, as we always have a device with access to Google within arms reach. We are more efficient task managers, information gathers, and communicators than our forebears ever could have been.
Yet, there is still a tremendous amount of friction; our thumbs are slow and clumsy, speech recognition hasn’t fully evolved yet. It is likely that this is one of the problems that AI will solve. With a neural lace embedded in our skulls, we will be able to communicate our thoughts and intentions directly to our devices. This “meaningful partial-brain interface,” Musk suggests is only four or five years away. In the meantime, here are several ways that marketers can use AI and machine learning to reduce friction and create engagement with our audiences.
7 Use Cases for Artificial Intelligence in Marketing
Content is one of the most likely and readily apparent applications of AI in marketing. Blog title generators have existed for several years. A natural and logical extension is for AI to write full blog posts, web pages, and email messages. Wordsmith is a natural language generator already in use by The Associated Press for short news updates.
AI is also powering content curation efforts such as Netflix and Amazon’s recommendations. Spotify is also a big user of AI to provide recommendations and create playlists such as “Release Radar” which gives users a personalized playlist of new music every Friday based on what users have previously listened to. When consumer behavior is analyzed by AI, predictive analytics can yield the best possible next steps for consumers to take.
2. Website Design
There are already more than 1 billion websites, so it follows that almost every layout and style has been tested somewhere online. Building upon the knowledge of what works and what doesn’t, companies like Grid and Wix already offer AI tools to help design a new website.
Google’s search algorithm is now powered by RankBrain, a machine-learning system that processes and provides search results. Since, 2008 Bing too, in an effort to cut into Google’s search dominance has machine learning to process its results. Accordingly, “AI has helped Bing more than double its share of the search engine market (to 20%)” The first implication of machine learning in search is that black-hat SEO tactics, keyword stuffing, link farms, etc. no longer have a place in marketing.
Then, of course, there is also the emergence of Amazon Echo, Google Home, Apple’s Siri, and Microsoft’s Cortana. These digital assistants, as discussed earlier, use natural language processing to facilitate voice search.
4. Social Media
The social networks themselves are already among the most involved in the race for artificial general intelligence. Individuals are empowered with AI tools in choosing which content they want to see and marketers have the ability to use AI in choosing who to target with ads.
The most promising use case of AI in social media, however, is the opportunity for personalized one-to-one communication afforded by chatbots. We spoke at length about how chatbots, using natural language generation, can be used to replicate human conversation on platforms such as Facebook’s Messenger in the previous webinar in this series.
5. Ad Buying
If you have bid for ads on Google’s AdWords platform, then you are among the 63% already using AI tools, perhaps without knowing it. The bidding option is powered by AI to give you the lowest possible cost-per-click (CPC).
Brands like Harley-Davidson, Dole, and Cosabella have also reported ad-buying success with Albert, an AI-powered platform that enables autonomous targeting and media buying.
6. Fraud Prevention
Many organizations may find immediate benefit in AI’s ability to analyze user behavior to prevent fraud and data breaches. By analyzing typical credit card usage, for example, AI can notify consumers and financial institutions when behavior is outside the norm indicating that fraud is being committed.
AI also promises a great reduction in the difficulties inherent in choosing the right price for a product or service. Organizations like Uber use tools such as PerfectPrice to “anticipate and react to changes in demand and...market in real time.”
At SilverTech, several of our partners are piloting or have recently rolled out AI tools. Let’s take a look at a couple of examples.
HubSpot - Scaling Inbound
With the recent acquisition of machine learning company, Kemvi, HubSpot is using AI to scale Inbound. It is easy enough for a single marketer to build out a workflow for the download of a single piece of content, but for larger organizations with perhaps tens of thousands of content pieces, building personalized nurture tracks can be daunting, if not impossible. HubSpot is hoping that AI will eventually build these workflows and learn as it goes how to optimize them.
Among the AI tools already rolled out or at least in Beta are Content Strategy and Predictive Lead Scoring. From HubSpot, “Content Strategy uses machine learning to understand the themes that search engines associate your content with. You’ll see useful topic suggestions, and detailed metrics like competitiveness and relevancy that help you hone your SEO strategy in on the right content areas.”
Lead scoring is an essential part of successful Inbound marketing; knowing when to pass off leads to sales or continue to nurture can make or break an opportunity. HubSpot’s Predictive Lead Scoring sets out to eliminate some of the guesswork in manual lead scoring. Using machine learning, “Predictive Lead Scoring builds a custom model and scores every contact in your database, telling you who is most likely to buy.”
Our partners at Salesforce, the cloud-based customer relationship management (CRM) platform, have also invested in AI powered technology. Introduced at Dreamforce in 2016, Salesforce Einstein is “an intelligence capability built into the Salesforce platform and focused on delivering smarter customer relationship management.” The aim of Einstein is to empower every Salesforce user with the insights of a dedicated data analyst. Like HubSpot’s AI tools, Einstein is meant to scale customer relationships, “by making intelligent predictions and recommendations about your deals.”
Getting Started with AI - It’s Easier than You Think
Many marketers become uncomfortable around the topic of artificial intelligence. The fear is: AI tools will eliminate marketing jobs. And, though, yes, it is likely that marketing departments will be radically transformed over the next 2-5 years, there are great opportunities for those who embrace the tools and prepare for their adoption as early as today.
If you are looking to get started with AI (beyond what you may already be using), it’s important continue to read up on the topic. Beyond our introductory webinar, there are some great organizational and operational resources available at Topbots. For philosophical and ethical considerations, it is also worth bookmarking the Future of Life Institute. Hopefully, the links gathered throughout this presentation offer several different paths for you to start your own research.
As you research, it is worth considering how AI tools could enhance and amplify your marketing. Even if you are not ready to make the move today, it is likely that you will need to determine over the next few months whether or not you will be building or buying an AI solution. Many organizations will also be faced with the decision in the near future as to whether they need an internal AI officer in the C-suite. It may seem far-fetched but large enterprises are already making such moves.
The next step is to start small. Build a bot. Chatbots are a great point of entry into AI. An FAQ bot is a great place to start. Transform your FAQ section into a bot that your audience can engage with on a platform such as Facebook Messenger.
Another important step to take is to purchase and use, whether it’s at home or in the office, an Amazon Echo or a Google Home. The more familiar you can become with technologies that use natural language processing and generation, the better prepared you will be to market to your audience with them as your medium. Much of our job as marketers is one by staying a step ahead of consumers, and adoption rates are growing quickly for these devices. Add one to your holiday shopping list, if you don’t have one already.
Today’s marketers need to prepare for the AI revolution and understand how they can use it to enhance customer experiences in our increasingly digital world.
Ekaterina Walter, Forbes
To return to our concern about the radically changing marketing department, let’s close with 5 skills you will need in order to succeed in an AI powered industry. In the recent article, How to Keep Your Marketing Job Through the AI Revolution, Forbes contributor, Ekaterina Walter suggests you need the following skills:
- Problem Identifier
- Computational Thinker
- Renewed Creative
- Open-Minded Thinker
These skills will help you embrace the possibilities of AI while keeping your skillset valuable in the new marketing department. As we conclude our presentation, it’s worth focusing on number 4 for a moment. Creativity is the human trait furthest from AI’s reach. Creativity “involves not only a cognitive dimension (the generation of new ideas) but also motivation and emotion, and is closely linked to cultural context and personality.” It is unlikely that machines will be able to improve upon this unique combination of traits.
As we approach “meaningful partial-brain interface” by 2023, machines achieving human-levels of intelligence by 2029, and perhaps “Singularity” by 2049, it is important for your individual success, and that of your organization, that you embrace the near future of marketing and are creative in doing so.