By Julie Hall, Orlando Area Chapter
In an eye-opening breakout session, Martin Waxman, MCM, APR, owner of Martin Waxman Communications, discussed how artificial intelligence (AI), machine learning and big data present many opportunities and challenges for communications professionals.
While AI may seem like the makings of a distant science fiction future, we actually already interact with these systems daily. Google, Apple, Instagram, Facebook, Netflix and Twitter among others, all use AI to make their systems more effective and user-friendly.
Waxman explained that there are three key types of artificial intelligence, warning attendees that when we develop advanced AI risk scenarios, we have to use the tropes of science fiction to shift our mindset:
- Artificial Narrow Intelligence (ANI) – This technology is meant to complete a single task, only much better than humans can. Examples include IBM Watson, digital voice assistants like Alexa, self-driving cars, and the world’s first robot citizen (of Saudi Arabia), Sophia.
- Artificial General Intelligence (AGI) – When machines become more like us, they can transfer knowledge between tasks. The AI becomes smarter than we are or ever will be, and could become conscious. Examples include the robots in Westworld or Her.
- Artificial Super-intelligence (ASI) – This level of AI would be smarter than the sum total of all humanity, seeing patterns that we can’t imagine. We may not be able to control this type of AI and don’t know how it would behave. What happens if the ASI changes its goal and becomes our master? Waxman asks whether it is ethical to enslave a superintelligence to do our bidding?
While this seems to be the makings of sci-fi nightmares, the breakout session focused mostly on the first category, narrow AI, which has been around since the 1950s.
The key driver of AI technology is big data, which we generate today in staggering numbers. 2.5 quintillion bytes of data are generated each year, falling into three main types:
- Structured data areare organized and labeled, like in an Excel spreadsheet.
- Unstructured data are more complicated, like the relationship between words in a phrase or the data with an image.
- Semistructured data combines both structured and unstructured data. Twitter is a great example. User names, locations, number of followers, etc. are structured, whereas the content of tweets, images and gifs are unstructured data.
More data = better predictions = better results
How does artificial intelligence actually work? AI systems use natural language processing (NLP) and computer vision to understand what we say and see, and use machine learning to start to understand the world around them. Like a small child, AI systems learn by absorbing what we teach it and what it learns independently. The machine then uses either human-supervised or unsupervised learning to make a prediction based on the highest statistical probability that it will be right.
AI in action
Google’s voice search tool understands us, even when we don’t provide it with all the details of what we’re looking for in follow-up questions. Waxman demonstrated by asking Google who the Prime Minister of Canada is (Justin Trudeau) and then following up to ask, “how many children does he have?” Google inferred that he was still asking about Trudeau, even though Waxman only provided the pronoun “he.”
According to Waxman, “that’s the beginning of a beautiful relationship.” And relationships are what we do best in PR.
Ethical dilemmas with AI
Waxman explained that data are the AI equivalent of a good meal. If we provide “junk food” data, the resulting AI will have data bias. Some examples he provided were facial recognition glasses being tested in China that police will wear and use to apprehend suspects.
The recent Google Duplex demo that featured an AI calling a business to schedule a haircut was so lifelike that the person on the other end of the phone had no idea they were speaking with a robot. This opens a lot of ethical questions about AI transparency (following public criticism, Google Duplex will now identify itself as an AI).
Impact on PR and the workplace
This will have huge implications for PR professionals, especially in an era of “fake news.” As bots write more news stories, including sports recaps and reporting of routine financial statements as is happening today, can we believe that a person wrote an article or a machine? Even if this technology is only used for good, it could lead to job losses for journalists and communications professionals.
In fact, 73% of adults believe AI will reduce more jobs than it creates. Any job that includes repetitive tasks – including some of what we do as communications professional and marketers – are at risk, as machines can do those things cheaper and more efficiently.
Waxman cautions that the speed and scope of change is reminiscent of the advent of social media and left the group with four recommendations.
- Training – Half of Americans are looking to their employers to provide retraining to compete in an AI dominated environment. How is your organization tackling this?
- Reimagine PR learning – Data science, statistics, and basic coding classes should be included in PR educational programs. Break down the barrier between academia and industry to develop (academia) and test (industry) AI theories.
- Start the AI conversation – Position PR as the strategic leaders, asking tough questions of other departments, including “how will AI transform our culture and workplace?”
- Redefine our value & role – How does PR strategically fit into this space? What are our goals: Are we the explainers of AI, script writers for chat bots, etc.?
This all takes time and energy, but we can’t be complacent. We have to actively participate in the debate. While communicators don’t need to become coders, we should start thinking like a programmer, have basic data literacy, and start to understand what this means for our industry.
Resources to learn more:
- Follow #AIinPR and see @martinwaxman’s AI Twitter list
- Books recommendations
- Human + Machine by Paul Daugherty and H. James Wilson
- AI for Marketers by Christopher S. Penn
- The Big Nine by Amy Webb
