What Are AI Chatbots and Why Are They Popular? – MUO – MakeUseOf
Generative AI chatbots are popping up all over the place, but what’s behind the sudden boom?
It seems like every company is scrambling to develop its own AI chatbot these days. You've also probably seen existing ones integrated into websites, apps, and messaging platforms. But what exactly are AI chatbots, and why do they seem suddenly mainstream?
AI chatbots have been around for decades, but artificial intelligence has recently advanced enough to make them seem almost human.
The first AI chatbots popped up in the mid-1960s when computer scientists tried to build programs that could have genuine conversations. These were pretty rudimentary and rule-based, though–think Eliza, the psychotherapy bot that could respond but not hold an actual coherent discussion.
In the 1970s, AI chatbots got a bit smarter with pattern-matching techniques. PARRY, developed in 1972 by Stanford psychiatrist Kenneth Colby, simulated a person with paranoid schizophrenia using syntax and word patterns to formulate responses. It fooled some into thinking they were chatting with a real patient.
Jabberwacky, created in the late 1980s, took a more playful approach. It aimed to simulate natural human conversations, albeit still in a limited way with predetermined responses. Many internet users interacted casually with Jabberwacky in its early days.
The bigger leap came in the 2000s with machine learning and natural language processing. SmarterChild, launched in 2001, analyzed volumes of real-life chat logs to learn patterns. Developed by ActiveBuddy, it lived on AOL Instant Messenger offering music recommendations, answers, games, and conversational quips to millions of users.
Unlike scripted bots, SmarterChild could dynamically respond to most topics by matching user input to trained chat data. It marked major progress in AI chatbots sounding more human. Teens in the 2000s loved chatting with it, even forming para-social relationships!
Fast-forward to today, and AI chatbots are trained on massive text datasets using neural networks, deep learning, and large language models to craft new responses on the fly, have personalized conversations, and display empathy.
Major players include ChatGPT, Bard, BingAI, and Perplexity AI.
You have to admit; today's AI chatbots are pretty slick when conversing. Unlike the clunky bots from back in the day, these modern AI-powered chatters have some nifty features that allow them to keep up their end of a conversation like a real human. Let's look at some of them:
As we've established, AI chatbots have come a long way from the rigidly scripted bots of the past. And as artificial intelligence keep advancing, they're sure to get even chattier and more helpful. But for now, let's explore why these clever bots are blowing up so fast…
With the massive investments and training that they undergo, it's no wonder AI chatbots are blowing up! Their knowledge and conversation skills make them a great resource.
AI chatbots can dig through mountains of data to hook you up with what you need—whether it's a tricky homework question or help coming up with a recipe. And they can do this way faster than asking your friend to Google it for you.
Being skilled in communicating in different tones and contexts, they can adjust their tones and vocabulary to fit into conversations on topics ranging from music and pop culture to physics and human anatomy.
On top of being research geniuses, AI chatbots are great for practicing conversations in different languages. For example, if you're studying Spanish, you can chat casually about your day, discuss interests like music and travel, or even debate non-controversial topics, all in Spanish!
AI chatbots are masters at providing inspiration and feedback throughout the writing process. If you're struggling to develop ideas, they can draw upon their vast databases to suggest directions to take the piece. Beyond that, AI chatbots can rephrase awkward or unclear sentences with improved wording and flow. Or you can ask them to rewrite passages in a different tone or style to polish and vary your writing.
If you're a programmer or still learning the ropes, AI chatbots can explain coding concepts and best practices in plain terms, fill in knowledge gaps, and even recommend projects to practice and apply your skills. Over time, regularly collaborating with an AI chatbot (versed in programming) can help you become a more self-sufficient, versatile developer. So rather than banging your head against the wall over code, let a friendly AI chatbot help you quickly write, improve, and debug programs. It makes programming less solitary and frustrating.
Lastly, one of the coolest things AI chatbots can do nowadays is provide entertaining company. You can tell them about your day, joke around, get their takes on stuff…you know, shoot the breeze with these AI pals.
And for some friendly competition, you can challenge your AI chatbot to word games, trivia matches, and other text-based contests. Heads up, by the way, they've got more knowledge than any human!
There's just so much you can do with AI chatbots!
AI chatbots may be the talk of the town, but they aren't perfect.
One of the biggest challenges is the tendency to hallucinate or make up information they don't actually have knowledge about. AI chatbots will attempt to generate plausible-sounding but inaccurate or nonsensical responses when prompted with questions beyond their training data. This can be dangerous if users assume the chatbot's responses are factual. There have even been instances of generative AI chatbots inventing precedents for legal cases!
Off the back of AI hallucination is a lack of a consistent personality or viewpoint.
Human conversations normally reflect the individual's unique perspective and areas of expertise. But AI chatbots have been explicitly designed without an opinion to avoid causing offense. While this can be helpful, it makes them generic and predictable over time. No depth of character develops cumulatively during a conversation.
Some critics argue these AI chatbots lack grounding in the physical world. Unlike humans, AI systems cannot perceive objects, make observations, or interact with the environment. All of their knowledge comes second-hand from language data.
For example, a chatbot like ChatGPT may be able to create a healthy meal plan for you, but it certainly can't cook a meal! This severely limits their practical applicability compared to a human assistant.
Security and privacy risks around AI chatbots' data collection have also sparked controversies. Critics have raised concerns about logged private conversations and potentially inspected without sufficient user knowledge or consent.
And like any technology, AI chatbots carry risks of misuse in the wrong hands. Their ability to generate believable text means they could be used to spread misinformation, plagiarize content, engage in fraud, or produce harmful instructions if not governed carefully.
There's no denying it; AI chatbots like ChatGPT, Bard, and Bing AI have gone mainstream. Their ability to understand natural language, converse intelligently, and generate content on demand has captured public fascination.
For many, AI chatbots represent an exciting preview of how AI could profoundly enhance our digital experiences. But this enthusiasm is tempered by uncertainty. The average person still has questions and doubts about them. Can they replace human intelligence and jobs? Are they safe and free of bias?
As artificial intelligent chatbots continue proliferating, these questions and misconceptions remain.
Oluwademilade is an MUO staff writer focused on everything tech. Since 2019, he has contributed to several websites, and prominent media outlets and professionals in the tech space have acknowledged his writing. Besides writing, Oluwademilade likes to read, play the bass guitar, and travel.
