“…It will mean that 95% of what marketers use agencies, strategists, and creative professionals for today will easily, nearly instantly and at almost no cost be handled by the AI…” – Sam Altman, OpenAI CEO, in his book Our AI Journey.
Before we let Sam determine our fate, let's unpack a little about Artificial Intelligence. A lot has been written about AI, but we're just beginning to understand all that it can do, how we put it into practice, and how we manage its possibilities.
AI is not as new as you may think. Google has used AI in Search for almost ten years, starting with its RankBrain algorithm. That AI model assisted Google in understanding search intent. However, for many, AI did not become part of their universe until the end of 2022 and into 2023, when Open AI's ChatGPT was launched and visited by over 1.5 billion users. Since that time and the launch of similar products, ChatGPT visits have leveled off to 100 million weekly users.
Ask yourself if you're currently using AI, what for, and what tools. You may be surprised that you've been using AI for years and didn't know it.
Now, quickly becoming a mainstream tool, we see what we, as marketers, can do with it and how it can give those willing to try AI tools an advantage over the competition. With the proper training and experimentation, we're learning that effective input and prompts are the first steps in harnessing the benefits of AI models. For those hesitating, it may help to start with a bit of background.
What is AI?
Artificial Intelligence (AI) is a technology that simulates human intelligence to complete tasks using data input and algorithms. It is a term used for a variety of technologies with different capabilities and functions. Some AI systems learn from past experiences to complete tasks. They can also be machines that become more intelligent from a large data intake.
There are many different types of AI, which are defined by categories, initial branches, capabilities, and functionality—some of which overlap with their technological abilities.
Types of AI
Most AI experts group the types into two to three main categories.
- Narrow AI or Weak AI works on small tasks with limited abilities, like an internet search or playing a game. This category of AI is not flexible and doesn't fully understand context. It can't learn or make adjustments.
- General or Strong AI can autonomously complete new tasks due to human-like cognitive abilities. The technology can learn, adapt, and carry out tasks. This type of AI could lead to highly functioning assistants.
- Super-intelligent AI is a future possibility in which the technology goes beyond human abilities to problem-solve and be creative on its own. Theoretically, this system would exceed human abilities and intelligence and might include self-awareness.
Technology/Branches of AI
- Natural Language Processing (NLP) includes interpreting human language for use in tools like translation services and chatbots. NLP is also used in generative AI, allowing the technology to respond to conversations supporting personalization processes and customer service.
- Machine Learning (ML) is an algorithmic technology that learns from data. Taking information and creating patterns to learn from can lead to filters, image recognition, and NLP. This type of technology can create predictive AI to identify patterns from past experiences and predict future events. Predictive AI and Autonomous AI can assist in predicting future customer behavior, allowing for personalized experiences.
- Reinforcement Learning, also a type of machine learning, improves through experience, trial, and error. The system looks for optimal results through positive experiences and may include image processing and self-driving cars.
- Deep Learning is another machine learning process that evaluates large amounts of data through neural networks, which include image recognition and voice control.
- Robotics examples are Flippy and Chippy flipping burgers and cooking French fries in fast food restaurants like White Castle and McDonalds. Another example is CaliExpress, the world's first fully autonomous restaurant powered by AI.
- Computer Vision includes medical image analysis, surveillance, and facial recognition.
- Expert Systems will answer questions or problem-solve systems set up by rules.
NLP, ML, and possibly reinforcement and deep learning are the branches of AI that we will encounter most often. As you can see, several AI possibilities are still futuristic or occur in a world of science and mathematics that we are not.
Functional AI
There are four main types of Functional AI that fall under the different categories and branches of AI outlined above.
- Reactive machines are machine learning models. They don't have stored memory and are one-task-focused. They take large amounts of data and spit it out without learning. This is currently the most common functionality.
- Limited memory machines are trained on data from current and past scenarios to perform tasks and make predictions. This type of AI is more advanced than Reactive AI.
- Theory of mind is a theoretical type of AI. There are currently no real-world examples of this technology. In the future, the technology may understand thoughts and emotions, impacting how it responds to others.
- Self-awareness is a future theory that AI can have a sense of self. This function would be self-awareness and consciousness.
With this knowledge of AI's actual and future technologies, functions, and capabilities, you can see what areas to focus on for your purposes. Before you get started, you should also understand the benefits and challenges of Artificial Intelligence.
Benefits
According to a recent Klaviyo + Qualtrics AI trends report for eCommerce marketers, 41% of marketers believe AI will improve quality and accuracy. Only 28% focused on customer engagement. 44% strongly agree that AI helps improve CX faster than they can on their own.
The primary benefits of current AI are that it makes our lives "easier" by supporting productivity and collaboration and creating efficiencies, which inevitably allow for cost savings. AI offers more efficient task automation, analyzes large amounts of data more quickly and accurately, and offers real health benefits in the assistance of medical diagnosis and patient trial matches. In a healthcare setting, it may also allow physicians and staff to automate patient notes to allow doctors more time, and it can be used to prioritize the order of responses to patient portal messages based on keyword training.
In theory, AI will eventually produce better results than humans can quickly and accurately create. AI search engines like Google's Search Generative Experience (SGE), which rolled out as AI Overviews earlier this year, understand keyword intent, rank pages based on apparent intent, and learn from searchers' engagement with the results shown. The goal is to offer results that more accurately provide for specific needs.
Many brand leaders hope AI will benefit their customers, make employees' lives easier, and give back time—which means money. They want the technology to provide a better personalized experience and deliver improved customer service.
Challenges
Despite these benefits, there are many concerns about AI. Programs may lead to miscommunication, misinformation, and bias – leading to ethical issues, security and privacy concerns, and a loss of control. Humans teach AI, and any input from a human will include unconscious or conscious beliefs. AI tools can only amplify human biases, requiring us to consider what prompts and inputs to use, the outcome of the input, and the implications.
There is also a significant learning curve and obstacles to AI adoption. Depending on an organization's willingness and ability to communicate expectations and provide employee training, it can create an unsafe internal situation. This may contribute to internal confusion and disorder, causing quality control and legal and data risks.
When employees upload confidential and proprietary data in an open AI system, that data may become the results for an outside organization's AI projects. For this reason, many companies using AI are developing internal systems like intranets to protect their data. These proprietary AI tools, while more secure, also limit output due to the limited input.
Businesses need to create AI standard guidelines for their workforce, develop training programs based on those guidelines, and clarify when, who, and how AI should be used. AI will be a constantly evolving technology that companies must stay informed about, if not ahead of.
AI and Media Services
AI may be used for everything from media plan development to executing buys and analyzing results. It may help analyze past media data and audience targeting more quickly to recommend plan updates. AI tools may also find more efficiencies in buying to save clients money or allow for expanded ad reach.
This new opportunity of AI in marketing should trigger brands to ask their agency partners how they are using it, if they are, for their marketing services and learn more about any proprietary platforms before they are included in their media strategy.
One of Vision Media’s partners, The Trade Desk, uses Koa™, a “powerful predictive engine,” to maximize analysis, uncover audience insights, and process large amounts of data to identify trends. These benefits help advertisers save time and money and optimize campaigns.
Artificial Intelligence's Horizon
Many technology developers and business leaders are conflicted about their work and the path they will create. At a recent leadership event on the Future of AI, one healthcare executive on the panel described her relationship with AI as a "love-hate relationship." Some believe it is moving too fast and there are too many unknowns.
Over 30,000 people recently signed a petition created by the global non-profit Future Life Institute, AI experts, and AI-focused NGOs to pause large AI experiments, believing there are too many risks to society and humanity at the pace we're going. However, as the saying goes, the train has left the station, and it may be too late to turn it around or even stop it from moving forward.
AI as a tool will increasingly become part of most of our work. While it shouldn’t replace our strategic thinking, it will and has already begun to manage our tasks.
While Sam Altman's prediction for the future of marketers and their partners may be extreme, the world is changing fast, and those who embrace AI may be in a better position than those who pretend it's not happening.
Additional Sources:
https://www.acxiom.com/marketing-trends
https://www.coursera.org/articles/what-is-artificial-intelligence
Main Image Source: pexels-tara-winstead-8386440 Digital lines and hand.jpg
