If you’ve been paying attention to social media lately, you’ve noticed that everyone is talking about AI – or artificial intelligence. While these AI programs have been around for a while, their development speeds and capabilities are skyrocketing. What seemed impossible yesterday is possible today.
Why did AI become so popular?
Explaining AI’s popularity spike is easy. The buzz started with the release of DALLE-mini (or Craiyon), an AI model which generates low-quality images based on short text prompts. Due to the humorous nature of these images and the free, accessible online interface, everyone started to experiment and share memes.
DALLE-2 was the next evolution, a model developed by Open-AI, which is several times larger than Craiyon and can generate more realistic, detailed images. Many others, such as Midjourney and StableDiffusion, are available either for free or through payment. Each has its own vibe and niche.
Text generation snuck up on a lot of us. When ChatGPT, an AI-powered chatbot, was released in November of 2022, the detail and consistency of its responses astonished the internet. It can write articles or stories, answer factual or philosophical questions, and even write computer code! It’s also free, which means anyone can use it.
How does AI work?
These artificial intelligence programs work somewhat like our brains, utilizing a “neural network.” The neural network takes in information from the outside and searches for patterns, bouncing it back and forth between different “neurons.” It gets better at the specific task every time it does it, becoming faster and more accurate.
But how does this turn into text or image generation? Text generation models predict the next word in a sentence based on what came before it. The AI is trained on massive amounts of text from the internet, so it has a pretty good idea of how people write. Image generation is similar – the AI is trained on extensive datasets of images from all over the internet and can predict where pixels should go based on the prompt it receives from the user.
What are the ethical concerns with AI?
Many people have raised ethical concerns with these programs.
The visual art community is understandably skeptical about AI art, arguing that it isn’t real and detracts from art created by real people. They’re also angry that art is often included in training datasets without the artists’ knowledge or consent.
Teachers and professors are facing an entirely different problem – cheating. Since ChatGPT is adept at answering questions and writing essays, students have been making it complete their homework.
There are also concerns from writers that AI will take over their jobs.
What do you think? Are the risks worth the benefits of using AI?