Over the past few days, you may have noticed something new within Meta's apps, including Instagram, Messenger, and WhatsApp. It is a chatbot with artificial intelligence.
Within these apps, you can chat with Meta AI and enter questions and requests like “What's the weather like in New York this week?” or “Write a poem about two dogs that live in San Francisco.” The assistant said, “Corgi was short and had a very wide butt, and Lab was tall and had a slippery tongue.” I can think of an answer quickly. You can also tell Meta AI to generate photos, such as an illustration of a family watching fireworks.
This is Meta's response to OpenAI's ChatGPT, the chatbot that transformed the tech industry in 2022, and similar bots such as Google's Gemini and Microsoft's Bing AI. Meta bot's image generator also competes with AI image tools such as Adobe's Firefly, Midjourney, and DALL-E.
Unlike other chatbots and image generators, Meta's AI assistant is a free tool built into apps that billions of people use every day, introducing this type of artificial intelligence known as generative AI. This is the most aggressive effort to date by a major technology company. To the mainstream.
“We believe MetaAI is the most intelligent AI assistant at our disposal today,” CEO Mark Zuckerberg posted on Instagram Thursday.
The new bot invites you to “Ask Meta AI anything you want,” but after testing it for 6 days, my advice is to tread carefully. Many mistakes can be made when treating it as a search engine. For now, we can have a little fun. Its image generator will be a smart way to express yourself when chatting with friends.
A Meta spokesperson said the technology is new and, like other AI systems, does not always give accurate responses. There is currently no way to turn off Meta AI within the app.
Here's what doesn't work well and what works well with Meta's AI.
Not a search engine
Meta has announced a chatbot that will replace web searches. Groups of friends planning a trip can search for flights while chatting by typing MetaAI queries into the search bar at the top of Messenger or Instagram, the company said.
To be frank, please don't do this. Meta AI fails spectacularly at basic search queries like finding recipes, airfare, and weekend activities.
In response to my request to look up flights from New York to Colorado, the chatbot listed how to take public transportation from the Denver airport to downtown. And when I asked for flights from Oakland, California to Puerto Vallarta, Mexico, the bot listed flights from Seattle, San Francisco, and Los Angeles.
When I asked Meta AI to look up a recipe for baking Japanese milk bread, the bot created a common bread recipe that omitted the most important step: tangzhong (the art of cooking flour and milk into a paste). is created.
The AI ​​also created other basic information. When I asked for suggestions for a romantic weekend in Auckland, the list included a fictitious business. When I asked him to tell me about himself (journalist Brian Cheng), he told me that he accidentally mentioned The Verge, a technology blog he never wrote about, even though he worked for the New York Times. It was written.
Bing AI and Gemini, which are directly connected to Microsoft and Google's search engines, have performed well for this type of search task, but clicking links in a classic web search is still more efficient.
don't ask me to count
AI chatbots work by looking for patterns in how words are used together, similar to the predictive text systems on your phone that suggest words to complete sentences. They all struggle with numbers.
Naturally, Meta's assistant is bad at counting. If you ask him for a five-syllable word that starts with w, he will respond with a four-syllable word, “wonderfully.” If you ask for a four-syllable word that starts with “w,” you'll get the three-syllable word “wonderful.” Gemini and ChatGPT also fail these tests.
focus on the words
Like other chatbots, Meta performed better the more information it provided.
It was great at editing existing paragraphs. For example, if you typed a Meta AI paragraph that felt redundant and asked it to tighten it up, the chatbot trimmed out all the unnecessary words. When you asked the bot to improve a sentence written in the passive voice, the bot rewrote it in the active voice and added context. When I asked to remove jargon from a paragraph written by a tech blog, the highly technical terminology was rewritten in plain language.
This is a great study guide
Meta AI is great when working with existing text, which helps with learning. For example, if you are in a history class and studying about World War II, you can paste his website with information about the war into the search bar and ask the bot to quiz you. The chatbot reads the information on her website and generates a multiple-choice test.
Can be used as fun emojis
The most appealing aspect of Meta AI is its ability to generate images by typing “/imagine” followed by a description of the image you want. For example, if you type “Imagine a picture of a cat sleeping on a window sill,” it will generate a convincing image in seconds.
Meta's AI is much faster than other image generation tools such as Midjourney (which can take over a minute). The results can be very strange, with images of people missing limbs or cross-eyed.
Ethics experts have expressed concern about the impact of creating fake images, saying they could lead to the spread of misinformation online. But in the context of using AI while chatting with friends and family on WhatsApp and Messenger, Meta AI is a great example of how generating fake images can be fun and safe when treated as a new form of emoji. .
In a group conversation with my in-laws, I mentioned that I was looking to buy a stroller that would be sturdy enough to withstand the winding roads in my neighborhood. In seconds, her wife used Meta AI to generate an image of a stroller with giant wheels that resembled her monster truck. The image was conveniently labeled “Imagined with AI.”