5 Generative AI Chatbots With Features You’ll Love

You may find these bots in places on the internet that you already frequent

Edward C. Baig,

AARP

Comments

En español

Published September 29, 2023

/ Updated December 27, 2023

GETTY IMAGES

ChatGPT exploded onto the scene a little more than a year ago to become the fastest-growing consumer application in history, suggesting a new future of tech had arrived and was not hyperbole.

San Francisco start-up OpenAI’s generative AI chatbot churns out computer code, letters, party ideas, poems, reports and scripts in seconds after typing a prompt. It became the poster child behind arguably the biggest change in computing since the internet, an artificial intelligence-fueled modern era.

Types of artificial intelligence

These aren’t the only types of artificial intelligence but can give you a glimpse of what’s out there.

Conversational AI. Advanced technology that allows computer programs to understand and respond to questions in everyday language, something that tech experts call natural language processing. This type of artificial intelligence includes chatbots and virtual assistants. Both generative and predictive AI are conversational.

Generative AI. A computer program that creates original content based on patterns of data it has accumulated rather than giving answers from something it already knows.

Predictive AI. This was the traditional artificial intelligence around before generative AI. It has the capability to learn from data and make decisions, predict next steps or solve problems based on that data.

Rule-based chatbots. These computer programs, which you’ll often encounter in customer service settings on websites, use scripts and information on a particular topic to generate answers to questions. At its most basic, if a customer asks a specific query, the bot responds with an answer already in its memory.

Although developments in artificial intelligence have been taking place since the mid-1950s, it took decades before the technology captured the public’s fascination with twin triumphs. In October 2011, a month after IBM’s Watson supercomputer defeated beloved Jeopardy! champion Ken Jennings, Apple added its Siri digital assistant to iPhones.

These days, OpenAI isn’t the only company exploring and exploiting AI technology on the front lines. Generative AI chatbots are trained on what techies refer to as large language models (LLMs), mammoth sets of data culled from the internet.

It’s worth emphasizing that these remain early times for AI, even though the technology was born 67 years ago. Chatbots are still beset by incomplete data and inaccuracies. They often spit out stuff that sounds plausible but is wrong or misleading, commonly known as “hallucinations.”

Given the daily stream of headlines, you need a scorecard to keep track of what every company is doing in the space or, at the very least, wish you could unleash an AI bot to help you stay on top of the topic.

Here’s a summary of what some leading AI companies have been up to lately and how you might engage with AI now or in the near future.

1. ChatGPT now knows some current events, responds to voice

ChatGPT rolled out Nov. 30, 2022, as mostly text-based, meaning you had to type a prompt into a computer and wait, usually not long, for a result.

On Sept. 25 of this year, OpenAI began bringing out fresh voice and image capabilities for ChatGPT. But for now, you must either be a business customer or subscribe to a $20-a-month ChatGPT Plus plan.

If you do, you can snap a picture of a landmark while sightseeing and have a live conversation with the bot about what you’re looking at. You might take a picture of your refrigerator or pantry to get ideas for dinner, something you might previously have needed to Google, tried to consult a recipe app or rummaged through your own cookbooks for inspiration.

You’ll also be able to have a back-and-forth conversation with ChatGPT, maybe to settle an argument or read a bedtime story to your kid or grandkid.

In November, OpenAI introduced customizable versions of its AIs called GPTs, available for now to Plus subscribers. Each GPT is tailored for a specific purpose.

You might consult a Sous Chef GPT for recipe suggestions based on the ingredients in your pantry or fridge. Math Mentor can help you work with youngsters with math homework. And Laundry Buddy can handle queries such as “How do I remove a coffee stain?” or “What does this label mean?” referring to an image of a clothing label you upload.

OpenAI also lets you build your own custom GPT.

Until recently, ChatGPT had only limited knowledge of global events into early 2022. That’s still the case for the free version of ChatGPT, which anyone can try by visiting openai.com or by downloading its app onto your smartphone from the Apple App Store or Google Play store.

That constraint doesn't fully go away for Plus and business customers; the last update of consequence came in April 2023. But the Plus crowd can use ChatGPT-4 to browse the internet with the help of Microsoft’s Bing search engine.

When asked, “What kind of reviews did the Leonard Bernstein biopic Maestro get?” ChatGPT responded that the movie received “positive feedback,” and supplied brief references to critic websites such as RogerEbert.com and RottenTomatoes.

And when asked, “What is the latest news on COVID-19 vaccines?” ChatGPT, with an assist from Bing, mentioned among other things a Centers for Disease Control and Prevention (CDC) alert “emphasizing the urgent need to increase vaccination rates for COVID-19, flu and RSV [respiratory syncytial virus],” as well as ongoing research on inhaled COVID-19 vaccines. Just under the response is an important reminder: “ChatGPT can make mistakes.”

To try some of the latest features as a Plus subscriber, click the menu with your picture, and then click Settings & Beta | Beta Features | Browse with Bing.

2. Google Bard has option to check via traditional search

By most accounts, Bard has had to play catch-up with ChatGPT. Google has started to leverage its AI technology by connecting its chatbot through what are called Bard Extensions to a suite of Google apps and services, including Gmail, Maps and YouTube.

You can ask Bard to summarize recent emails from your kid’s school or create an itinerary for a family visit to Portugal. It also can summarize and explain documents stored in Google Docs or Google Drive.

To get going, head to bard.google.com; make sure you connect Bard to your Google Workspace. Meanwhile, if you’re unsatisfied or not sure an AI response generated by Bard is accurate, click on the “G” logo to double-check the answer in search.

If you’re worried about privacy, Google claims it does not collect or store any personal information about its users and that it uses automated tools to help remove email addresses, phone numbers and other personal information before human evaluators eventually review Bard’s work. The subset of conversations shared with humans is not associated with user accounts, Google says.

In early December, Google squarely took aim at ChatGPT by showcasing Gemini, a powerful “multimodal” software model that drives Bard and other AI ventures at Google. Multimodal means the model can process a mix of inputs, including text, audio, computer code, physical objects and visuals.

At one point during an impressive attention-grabbing, six-minute promotional video, the AI system was shown an orange next to what appeared to be a chocolate-chip cookie. The objects were properly recognized as food and the AI piped in that the fruit was the healthier choice.

But Google faced criticism when it was discovered that the video was edited and not captured in real time, giving the impression that Gemini is faster and further along than it is now.

Google has opened up Gemini Pro, the middle of three sizes and the one showcased in the video, to developers. Versions of the smallest size, Gemini Nano, are just starting to power certain features in Google’s flagship Pixel 8 Pro smartphone, including the ability to summarize recordings in the Recorder app.

Video: 4 Tips to Get the Most out of AI Tools

3. Microsoft has new Bing and Copilot, a new digital assistant

As an investor in OpenAI, Microsoft leverages technology in ChatGPT inside its own Bing search engine. The new Bing, as it is referred to, can display regular search listings next to AI-generated chat exchanges.

Meanwhile, Microsoft is pushing a new AI companion it calls Copilot, made available free as part of the Windows 11 update for PCs and in an update to the Microsoft Edge browser, Microsoft 365 and Bing itself. You’ll be able to launch Copilot from the taskbar in Windows or via a Win+C keyboard shortcut.

Microsoft recently retired its Cortana app, which had been its version of the Alexa and Google Assistant digital assistants with nowhere near the popularity. Copilot now assumes this assistant role.

You might ask it to organize your windows on a PC or summarize articles you’ve read. Microsoft points to other possible tasks: Ask Copilot to “Write a paragraph about minimalist architecture,” or “Create an Instagram reel with photos from my vacation.”

Separately, Microsoft is bringing the DALL-E AI text-to-image generator, also from OpenAI, to its Paint program in Windows.

What’s more, Microsoft just announced a new capability for Copilot you can try now: The ability to turn a single prompt into a brand new song, complete with lyrics, vocals and background music. Microsoft is partnering with an AI-based music creation start-up in Cambridge, Massachusetts, called Suno.

I asked Copilot to create a new song that high school kids could perform at a college audition. When I wasn’t wild about the first 36-second song the AI produced, I asked it to try again with more of a rock feel. The new song was a bit longer and slightly better.

Google and Facebook also have been testing music creation tools based on generative AI.

4. Amazon Alexa brings next generation of AI to your home

Lots of people ask Amazon’s popular Alexa voice assistant inside Echo smart speakers and other devices to play a song, set timers or tell them today’s weather. At a September press event in Northern Virginia, Amazon outlined ways in which you’ll be able to converse with Alexa more naturally, based on Amazon’s own AlexaLLM.

When Alexa-capable gear is embedded in a smart home, you might say something like, “Alexa, it’s chilly in here” and Alexa will know to raise the temperature on your smart thermostat. You won’t need to spell out “Raise the thermostat to 73 degrees.”

You might ask Alexa: “Turn on the TV, make the living room lights look like the aurora borealis and clean the dining room,” and Alexa will oblige, based on the other smart devices linked to it. Or before you go to bed, you can ask, “Alexa, is everything locked up for the night?”

Through an update that will be coming to Amazon’s Fire TVs and streaming sticks, Alexa might also help you find something to watch with AI-based queries like these:

“Find movies for date night.”
“Find movies we can sing along to.”
“Find movies that should have won an Oscar.”
“Find movies that are free to me or that I haven’t seen yet.”

One unsettling development to take note of: Now departed Amazon executive Dave Limp told Bloomberg that future versions of Alexa could eventually require a subscription.

5. Meta’s AI takes on various personalities with expertise

The Facebook parent and company behind Instagram, Messenger and WhatsApp has launched a conversational rival to ChatGPT called Meta AI, but is taking a far more playful approach than others.

Your designated Meta AI can help you search the web with an assist from Microsoft Bing, generate images from what you describe or provide recommendations on places to eat or visit.

But here’s where it differs: Many of the Meta AI bots you choose to engage with are “played by” and represented by real-life celebrities who assume different personas and specialties. Among them: retired NFL quarterback Tom Brady in the role of Bru, the “sports brain”; tennis star Naomi Osaka is Tamika, a “manga master”; Paris Hilton is Amber the “detective”; and Snoop Dogg is Dungeon Master, the “guide to games.”

The New York Times conducted a text conversation in WhatsApp with another of Meta’s AIs, a 19th century representation of British author Jane Austen. You’ll engage with these and other AI assistants through one-on-one or group chats inside Meta’s various apps, as well as new versions of Ray-Ban Meta smart glasses and Quest 3 augmented reality headsets.

Meta’s approach recalls another AI company, Character Technologies, which lets you design and text with living and dead characters from William Shakespeare to Oprah Winfrey. Sure, the Character.AI conversation is made up, but advances in AI are real and coming fast.

This story, originally published September 29, 2023, has been updated to include fresh AI developments.

%{postComment}%

Edward C. Baig covers technology and other consumer topics. He previously worked for USA Today, BusinessWeek, U.S. News & World Report and Fortune, and is author of Macs for Dummies and coauthor of iPhone for Dummies and iPad for Dummies. Follow him on LinkedIn; Threads; and X, formerly known as Twitter.