Artificial Intelligence: Past, Present, & Future
Hey Siri, what’s the weather like today? and Alexa, turn off living room lights are voice commands that we are all familiar with — maybe even ones that we’ve used ourselves. In today’s day and age, artificial intelligence (AI) seeps into our everyday lives. In our smart homes, we can rely on our AI-powered assistants to help us maintain the temperature of our houses based on our patterns and preferences. Not to mention, the growing presence of smart cars, like Tesla, where AI is integrated into its systems to allow virtually self-driving capabilities. AI even finds its way into our social media platforms and streaming services using machine learning algorithms to personalize content based on our activity. AI is everywhere and it is constantly evolving. In just these past few years, ChatGPT was introduced, an AI-powered chatbot that allows users to engage in humanlike conversations that answer questions and generate content. Whether you’re looking for homework help or ideas to aid your writer’s block, ChatGPT’s range of topics of conversation is endless. So how has our understanding and perception of AI changed over the years?
Well, despite AI’s growing prominence in our daily lives, relative to the 100 top-viewed Wikipedia pages that average upwards of 50,000 page views daily, AI does not appear to be a popularly searched subject on Wikipedia given its average of about 15,000 daily page views(Pageviews Analysis). However, taking into account the “AI effect” may suggest that once an application of AI becomes practical and common enough it is no longer considered AI (Wikipedia). AI is divided into two main categories: weak and strong AI; where weak AI is a type of AI that is designed to solve a narrow set of problems and strong AI typically has attributes of artificial consciousness and superintelligence (Matthew Renze). Examples of weak AI include digital assistants, like Siri or Alexa, which mainly rely on basic technology like voice recognition and the Internet to function. Weak AI also encompasses Roomba or Tesla vehicles, both of which appear to have autonomous capabilities, like self-vacuuming and self-driving respectively, yet still require human intervention and do not possess general human-like intelligence. Given all these inventions are examples of weak AI, it makes sense why page views for Siri, Alexa, Roomba, and Tesla are all generally below AI.
Before the growing presence of ChatGPT, when AI comes to mind, most people may think of sci-fi movies where robots are taking over mankind. Such extreme examples of AI are likely explained by the intrigue of strong AI, where examples of weak AI have become more normalized in our daily lives. As shown in the visualization below, it is only really ChatGPT that significantly exceeds the overall page views of AI. Even though ChatGPT and other AI chatbots are still considered weak AI, ChatGPT’s capabilities of generating content and human-like conversations coupled with its free access, ease of use, and community-driven growth have captivated individuals everywhere (MakeUsOf).
We can also understand how society's understanding and perception of AI has changed over time by looking at how the length of AI’s Wikipedia page has grown since 2002. As shown in the visualization below, roughly between 2002 to 2021, the length of the article has overall steadily increased in size. AI as a concept has been ever-changing and evolving since Antiquity (Wikipedia). In just the past 20 years or so, there have been many notable milestones that have contributed to our understanding of AI. To name a few, the Defense Advanced Research Projects Agency (DARPA) held the 2007 Urban Challenge for driverless cars to compete to obey traffic rules and operate in an urban setting. In 2012, Apple first launched Siri on the iPhone 4s and Amazon’s Alexa followed soon after. More recently, OpenAI first launched GPT-3, a large language model that has “digested billions of words to become well-versed in understanding human language, analyzing the meaning behind words, and generating language independently” (HubSpot). In 2022, OpenAI launched ChatGPT, different from GPT-3 in the sense that “while GPT-3 is a general-purpose language model that can be used for a wide range of NLP tasks, ChatGPT is designed to excel at generating human-like responses in a conversation” (Jaclyn Curtis).
It’s no wonder that the launching of GPT-3 caused a spike in AI’s Wikipedia article length as this new concept entailed adding a boatload of information. What’s interesting is the dip in article length following the launch of GPT-3. One possible explanation is that with GPT-3’s launch, Wikipedia’s page for OpenAI became more established. With OpenAI rolling out launches of GPT-3, ChatGPT, and its successive versions, information about the organization’s history, motives, strategies, and more began to populate its page.
All in all, what we know about AI is constantly growing. AI lies within our daily lives, yet also remains a thing of the future we are always trying to pursue. We should all be waiting on the edge for what’s to be added next to Wikipedia’s AI page. What do you think the next AI innovation is?
