Welcome to readin – the best world tech news chanel.

Meta's AI assistant is fun to use, but can't be trusted| GuyWhoKnowsThings


Over the past few days, you may have noticed something new within Meta apps, including Instagram, Messenger, and WhatsApp: a chatbot with artificial intelligence.

Within those apps, you can chat with Meta AI and type questions and requests like “What's the weather this week in New York?” or “Write a poem about two dogs who live in San Francisco.” The assistant will immediately give answers, such as “The corgi was short, with such a wide butt, the lab was tall, with a tongue that slithered.” You can also tell Meta AI to produce images, such as an illustration of a family watching fireworks.

This is Meta's response to OpenAI ChatGPT, the chatbot that revolutionized the tech industry in 2022, and similar bots, including Google's Gemini and Microsoft's Bing AI. The Meta bot image generator also competes with AI image tools like Firefly, Midjourney and DALL-E from Adobe.

Unlike other chatbots and image generators, Meta's AI assistant is a free, in-app tool that billions of people use every day, making it the most aggressive push yet from a major technology company. technology to bring this type of artificial intelligence, known as generative AI. to the mainstream.

“We believe Meta AI is now the smartest AI assistant you can freely use,” Mark Zuckerberg, the company's chief executive, wrote on Instagram on Thursday.

The new bot invites you to “ask Meta AI anything,” but my advice, after testing it for six days, is to approach it with caution. It makes a lot of mistakes when you treat it like a search engine. For now, you can have fun: its image generator can be a smart way to express yourself when chatting with friends.

A Meta spokeswoman said that because the technology was new, it may not always provide accurate answers, similar to other artificial intelligence systems. There is currently no way to disable Meta AI within apps.

Here's what doesn't work well (and what does work) in Meta's AI.

Meta announced its chatbot as a replacement for web search. By typing queries about Meta AI into the search bar at the top of Messenger or Instagram, a group of friends planning a trip could search for flights while chatting, the company said.

I'll be frank: don't do this. Meta AI fails spectacularly on basic search queries like searching for recipes, airfares, and weekend activities.

In response to my request to search for flights from New York to Colorado, the chatbot listed instructions on how to take public transportation from the Denver airport to downtown. And when I asked about flights from Oakland, California, to Puerto Vallarta in Mexico, the bot listed flights leaving from Seattle, San Francisco and Los Angeles.

When I asked Meta AI to find a recipe for baking Japanese milk bread, the robot produced a generic bread recipe that skipped the most important step: tangzhong, the technique of cooking flour and milk into a paste.

The AI ​​also created other basic information. When I asked him for suggestions for a romantic weekend in Oakland, his list included a fictional business. And when I asked him to tell me about me (journalist Brian Chen), he said he worked at The New York Times but incorrectly mentioned a technology blog I never wrote for, The Verge.

Bing AI and Gemini, which are connected directly to Microsoft and Google search engines, performed better in these types of search tasks, but clicking on a link through a traditional web search is even more efficient.

AI chatbots work by looking for patterns in how words are used together, similar to the predictive text systems on our phones that suggest words to complete a sentence. They have all struggled with numbers.

As expected, Meta's assistant sucks at counting. When you ask her for a five-syllable word that begins with the letter w, she will respond with “wonderfully,” which has four syllables. When he asks her for a four-syllable word that begins with w, she will offer “wonderful,” which has three syllables. Gemini and ChatGPT also fail these tests.

Like other chatbots, Meta worked better the more information you gave it.

He excelled at editing existing paragraphs. For example, when I sent Meta AI paragraphs that looked verbose and asked for the paragraphs to be adjusted, the chatbot removed all the unnecessary words. When I asked it to improve a sentence written in the passive voice, the bot rewrote it in the active voice and added more context. When I asked him to remove jargon from a paragraph written on a technology blog, he rewrote highly technical terms into plain language.

Because Meta AI is best when working with existing text, it can be useful for studying. For example, if you are taking a history class and studying World War II, you can paste a website with information about the war into the search bar and then ask the robot to give you a quiz. The chatbot will read the information from the website and generate a multiple choice test.

The most attractive aspect of Meta AI is its ability to generate images by typing “/imagine” followed by a description of the desired image. For example, “/imagine a photograph of a cat sleeping on a windowsill” will produce a convincing image in a few seconds:

Meta's AI is much faster than other image generators like Midjourney, which can take more than a minute. The results can be very strange: images of people who occasionally lacked limbs or appeared cross-eyed.

Ethicists have expressed concern about the implications of generating fake images because they can contribute 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 positive example of how generating fake images can be fun (and safe) if we treat it as a new form of emoji.

In a group conversation with my in-laws, I mentioned that I was shopping for a sturdy baby stroller that could withstand the crooked roads in my neighborhood. Within seconds, my wife used Meta AI to generate an image of a stroller with huge wheels that made it look like a monster truck, emblazoned with a helpful label that read, “Imagined with AI.”


Share this article:
you may also like
Next magazine you need
most popular

what you need to know

in your inbox every morning