You may not know it, but you’ve probably benefited from some recent advances in artificial intelligence, or AI.
AI powers your Siri and Alexa, those cheerful personal assistants that answer your questions about the weather or who won the big game last night. AI controls self-driving cars and the maps we use to navigate when we take a trip. When you call customer service for your bank or credit card, you may initially speak with an AI model of a human, trained to answer basic questions. And it’s AI that recommends the Netflix movies you’ll want to watch.
AI has been hiding in plain sight, helping make all kinds of technology smarter. But now AI is taking the spotlight, with a wave of new technologies, called generative AI, that generate content based upon your instructions.
Let’s see what these new tools can do.
The Chatbots: Chat GPT and Google Bard
Chatbots have been built into other applications, like online customer service, for some time. But Chat GPT (CHAT Generative Pre-Trained Transformer) was the first multipurpose commercial chatbot that end users could control. It made waves when it was introduced by Open AI in late 2022 as a free tool. It quickly garnered more than 100 million sign-ups.
Chatbots mimic conversation through a text window: you ask a question, and it replies. It’s like doing an online search, but instead of getting pages of possible relevant results, you get one well-written answer.
How does it know what to say? Open AI — a California-based artificial intelligence research lab — trained Chat GPT on the public internet, using machine learning to help it understand context and relationships across related pages. It has digested mountains of online information, but it does have limitations. It can make mistakes, so you should be wary and check its answers. Indeed, the internet is filled with funny and not-so-funny examples of AI devices not performing as desired. And it doesn’t know anything that happened after it was trained in 2021, so don’t go there looking for last night’s baseball scores.
I typed this question: “Is it better to take Social Security at full retirement age or at age 70?” and it generated a thoughtful, complete and well-written answer. If you did a Google search and synthesized the top five to ten results, you might end up with the same approximate guidance.
I then asked if it’s better to use shortening or butter when making pie crust. Its 233-word reply emphasized the superior flavor of butter but the lighter texture of shortening. This is consistent with my experience. Testing Chat GPT further, I asked it to tell me a joke. Its reply: Why don’t scientists trust atoms? Because they make up everything!
The “chat” part means that the software is aware of the previous interactions. So, in the example above about retirement age, I asked it to regenerate the answer as a limerick. In a flash, it replied:
In choosing Social Security’s sway,
Options beckon to come or to stay.
Full retirement or seventy,
Decisions can be weighty,
Consider your needs and you’ll find your own way!
Chat GPT has also been integrated with the Bing search engine, so you can access its features there without having to create an account. Unfortunately, to use the service, Bing forces you to download and use its Edge browser.
Not to be outdone, Google — another California-based technology company — has released its competing Bard service. And since Google has already indexed most of the web for search, you would expect it to be a leader in AI. Overall, Bard performs much the same as Chat GPT. But there are some differences. Chat GPT is a more capable writer. Bard tends to spit out chunks of information instead of a single, continuous narrative. Bard provides links in its responses, which can be useful if you want more information. And, like Google’s personal assistant speakers, Bard can be used to help book a vacation or assist with meal planning. Significantly, Bard, like Google, is always up to date, so it can help with news-related questions.
Photo magic with AI
If you’re a photo buff, there are some useful AI-based tools to help you improve your images. These tools can remove or replace backgrounds, smooth your skin or do a range of image adjustments. If you’re a professional, look for the AI plug-in for Adobe Photoshop. To access a limited version for free, visit the Adobe Firefly website.
Another jaw-dropping skill that AI is capable of is the creation of images from text descriptions. DALL-E, another AI tool from Open AI, is a leading solution. I asked DALL-E to generate images from this prompt: “children joyfully play at a splash pad on a hot summer day.” It generated four different versions of this scenario, and then offered to do variations on each.
Ethics, anyone?
These AI tools are both exciting and disturbing. Already, students are using Chat GPT to write essays for their homework assignments – and you don’t have to be a teacher to know that’s wrong. But other questions beckon: who is the author of AI-generated content? Can it be copyrighted? If so, who owns the copyright? After all, AI-generated text comes from a massive online database, much of which is privately owned by others. As AI becomes more powerful and more widely used in our culture, expect debates and possibly new laws that affect its use.
There have also been warnings from reputable sources that AI, because of its ability to evolve rapidly on its own, poses a threat to society and even the very existence of humankind if left unchecked — the kind of scenario science fiction writers portrayed decades ago. But let’s leave that discussion for another time.
As you use information technology in your daily life, be on the lookout for AI-enhanced tools. And if you’d like to explore some of these technologies, here are some links to help you get started:
Chat GPT and DALL-E, both available at: https://platform.openai.com/apps. Chat GPT is free; DALL-E requires that you purchase credits. A more advanced tool, Chat GTP4, is $20 a month.
Google Bard, free at https://bard.google.com
Adobe Firefly, free at https://firefly.adobe.com
Bing AI-assisted search, free at https://www.bing.com/new
David Kamerer is a longtime Wichita resident and early adopter of technology who teaches at Loyola University in Chicago.