Meta plans to implement technology to detect AI-created images
Meta, the parent company of Facebook and Instagram, plans to implement technology to detect and label images created by AI.
unbranded – Newsworthy
All the big tech companies are trying to one-up each other with “Hey, look what our artificial intelligence can do now” updates, and on Thursday, Facebook’s parent company Meta rolled out its latest update. updated in a version called Llama 3.
Unlike previous versions of Llama, Llama 3 has its own website (meta.ai), where you can use the tool for free. It’s a riff on the ChatGPT formula – no fees, go talk to him and see what he can do for you. The problem, of course, is that you log into it with a Facebook account.
“Big day here. We are launching the new version of Meta AI, our assistant, on which you can ask any question through our applications and our glasses,” Mark Zuckerberg, CEO of the meta-platform said in a statement sent to USA TODAY. “We’ve integrated it into the search box right at the top of WhatsApp, Instagram, Facebook and Messenger, so whenever you have a question, you can just ask it here. We’re also launching a number of unique authoring features. Meta AI now creates animations and creates high-quality images so quickly that it generates and updates images for you in real time as you type.
Some of my friends already see the “Ask Meta AI Anything” prompt at the top of their search feeds, but it hasn’t appeared in any of mine yet. However, meta upgrades are typically deployed this way; some people receive them immediately, while others have to wait a bit.
Will this change your life in any way? Absolutely not. Do you care? Maybe. Let’s go.
What does meta-AI do?
The coolest thing I’ve seen it do so far is generate animated characters and realistic images that update while you’re still typing the prompts.
Just type what you want to see and the image will start generating as you type. Here’s an example: Do you want to see a girl with curly hair steal a dragonfly? Type in “curly girl” and there it is. As you can see in the example above, the image continues to adapt to fit your story as you type. The only limit is your imagination. It’s like you can choose your own adventure, but without scary consequences.
What’s really cool about real-time image generation is that it doesn’t wait for the entire prompt to be written. Instead, it begins to “see” each object as you write about it and adjusts the final image in real time. This requires powerful hardware and software that are fascinating to watch in action.
What could go wrong?
This is a huge waste of time and obviously a very powerful tool. We’ll have to see if Meta’s obscene content filters are up to the task, though, as it’s a problem that has plagued similar tools from Microsoft, Midjourney, and everyone else.
As for guardrails, most top AI image generators from Microsoft and others have blacklisted keywords that allow them to generate images of famous people like stars of cinema and high-level politicians. The datasets used by these models generally don’t include “normal people,” so typing your friend’s name into the AI won’t produce anything that looks like them.
However, there are always ways to circumvent these safeguards. Dedicated users will likely figure out how to get certain celebrities to appear in images or even create NSFW (not safe for work) content using the system, something every other AI image generator has been fighting against since day one .
Meta AI to find and win arguments with friends
Why not generate fake images that people think are real, like those of politicians in the upcoming election? I’ll get to that in a second, but first, here are a few other things the company claims the new Meta AI can do:
◾ Act as a personal assistant. “We believe Meta AI is now the smartest AI assistant you can use for free – and it’s available in more countries through our apps to help you plan your dinner based on what’s happening in your fridge, to study for your exam and much more. »wrote the company in a press release. I see this as a way to win arguments with people on apps and in real life.
◾ Enable more powerful search. You can now use Meta AI “in feeds, chats, search, and more in our apps to get things done and access information in real time, without leaving the app you’re using,” the company says.
◾ Create images faster and more personalized. “You can create album art for your band, decorating ideas for your apartment, personalized animated GIFs and much more.”
During a video call demonstration earlier this week, a Facebook spokesperson said it expects people to use Meta AI “to have fun and be more social.”
To me this means that they are trying to make Facebook more fun for not just our parents and to make Instagram more interesting so that young people don’t continue to abandon it for TikTok. (Hard, but somewhat true, even though Facebook still has about three billion monthly active users, or about a third of the world’s population, according to the digital marketing site. Social push.)
Llama 3 will be available in 12 countries, including Canada, Australia and South Africa, all of which are benefiting from Meta’s AI for the first time.
Is meta-AI safe?
Llama 3 is the next generation of Meta’s great open source language model. It is similar to ChatGPTwhich is about to unveil its improved model, GPT-5. Models like these power AI-driven applications like chatbots, image generators, and virtual assistants.
Upgrading from Llama 2 (launched in July 2023) to Lama 3 should means more accurate responses, more human conversations, and a broader knowledge base for the AI applications that use it. Meta wasn’t the first to launch a super-powerful AI model – that honor, of course, went to OpenAI with ChatGPT – but it has been working feverishly to catch up. Great language models like ChatGPT and Llama are incredibly versatile and can be used for many things. They are the brains behind AI chatbots, developer coding tools, automated writing assistants, and many other tools that are becoming more ubiquitous by the day.
Building a large language model (LLM) means being able to power the next generation of mega-powerful user-friendly applications, and that’s something Meta is obviously interested in. It launched Llama 2 in July last year and its successor is already here, which tells you everything you need to know about the amount of resources Meta is devoting to the project.
The power of Llama 3 comes from the even larger AI database, which Meta has in abundance. The company tells me that Llama 3 draws on the vast stores of information generated by people who use the company’s products. However, it promises that personal and private information will not be transmitted to its language models.
That’s a comforting promise, but can they guarantee that no sensitive information ends up in any of these AI datasets? Not if history tells us anything, and that’s the cost of using social media.
Can a chatbot using Llama 3 technology be used for nefarious purposes? Can he pretend to be you or someone you know? Could he actually “deeply fake” your online identityallowing a scammer to tarnish your name or worse?
Meta claims that such a scenario is not possible thanks to various protection measures, but when we signed up with Facebook over a decade ago, we could never have guessed that one day our content would feed AI. Facing this reality might not sit well with many longtime Facebook users.
Specifically, Meta states that they will not let any “public figures” be used in generating images or videos. If you type in a person’s name, it might generate what a computer thinks a person with that name might look like, but that wouldn’t necessarily be accurate.
It also has several NSFW safeguards built in to protect against pornography, racism, and other issues that affect generative AI, but we have no idea how effective these guardrails are.
Why is Meta investing so much in AI?
Meta does this because having a large language model like Llama 3 means you can add new features to your apps, like adding AI to search features on Facebook and Instagram. This adds a layer of intelligence and can reduce friction between users and the apps they love.
What’s new with the Facebook app? : Why is the Facebook app different? Meta rolls out new full-screen video player
By building its own powerful AI, Meta can also sell licenses to companies who want to use it for their projects. ChatGPT operates on the same licensing business model, where access to AI is granted to businesses and their users for a specific fee. Meta stands to make a lot of money with an LLM in a variety of ways, including advertising (think auto-generated ads that interest you in particular) and selling access to this AI.
Facebook has one of the largest stores of data of any company, thanks to over a decade of user-generated content in the form of posts, comments, replies, and more. This data teaches Llama how humans speak and make connections between words and objects, places and concepts. The smarter it becomes, the more useful it should be.
That’s the theory, but I’ve been interested in technology long enough to be skeptical at best.
