These trends have practical implications for us, ordinary users. The estimates of the Massachusetts Institute of Technology (MIT) are worth listening to: their predictions for 2023 have come true almost completely, and current predictions, by the way, have already begun to come true.
Beyond the forecast, it is clear: large language models will continue to dominate, regulators will become bolder, and AI challenges will set the agenda for academics.
MIT (Massachusetts Institute of Technology) is one of the most prestigious technical educational institutions in the United States and the world. MIT occupies leading positions in prestigious rankings of universities in the world, is an innovator in the fields of robotics and artificial intelligence, and its educational engineering programs are recognized as the best in the United States from year to year. Here are the main trends in the development of artificial intelligence according to MIT.
Personal chatbots
In 2024, tech companies that have invested heavily in generative AI will have to prove that they can monetize their products. To do this, AI giants like Google and OpenAI are betting big on custom solutions. Both are developing user-friendly platforms that allow people to customize powerful language models and create their own mini chatbots for their specific needs, without any coding skills.
In 2024, generative AI could become truly useful for the average non-techie, and we’ll see more people working with millions of tiny AI models. AI models like GPT-4 or Gemini are multimodal, meaning they can process text, images, and even video. This could open up a whole new set of possibilities.
For example, a real estate agent can load text from their previous listings and set the model to generate similar text with the click of a button. Then, load videos and photos of new listings and ask AI to generate a description for their new property.
Of course, success depends on whether these models work reliably. Language models are often invented, and generative models are riddled with biases. They are easy to hack, especially if they can browse web pages. Tech companies have yet to solve either of these problems.
Video — the second wave of generative artificial intelligence
It’s amazing how quickly new things become familiar! The first generative models for image creation went mainstream in 2022. Images from OpenAI’s DALL-E, Stability AI’s Stable Diffusion, and Adobe’s Firefly have flooded the internet with images ranging from the Pope in Balenciaga to award-winning artwork.
the results were so-so, but the technology is improving rapidly. Runway, a startup that builds generative video models (and the company behind Stable Diffusion), releases new versions of its tools every few tunisia number data months. The latest Gen-2 model still generates videos that are only a few seconds long, but the quality is simply amazing. The best clips aren’t that far off from what Pixar makes.
Movie giants like Paramount and Disney are exploring the use of generative AI in their production processes. The technology is already being used for dubbing, and it’s opening up new possibilities for special effects.
Runway has organized an annual AI Film Festival, which showcases experimental films made using various AI tools. This year, the festival has a prize pool of $60,000, and the top 10 films will be screened in New York and Los Angeles.
In 2023, a deepfake of Harrison Ford will star in the film Indiana Jones and the Dial of Destiny. This is just the beginning.
Beyond the big screen, deepfake technology is also gaining popularity for marketing and educational purposes. For example, the UK-based Synthesia creates tools that can turn a single actor’s performance into an endless stream of deepfake avatars that repeat any script you give them with a single keystroke. The company says its technology is now used by 44% of Fortune 100 companies.
The ability to do so much with so little money poses serious questions for actors. Concerns about the use and misuse of AI by studios were at the heart of last year’s SAG-AFTRA strikes. But the technology’s true impact has yet to be discovered.
Electoral disinformation created by artificial intelligence will be everywhere