At an event in San Francisco in November 2023, Sam Altman, CEO of artificial intelligence company OpenAI, was asked what surprises the field would bring in 2024.
Altman responded that online chatbots like OpenAI's ChatGPT would make a "surprise leap."
James Manyika, a Google executive sitting next to him, nodded and said, "Second." ”
This year's AI industry will be defined by one main characteristic: as advancements continue to evolve, the rapid advancement of technology has made AI a reality. Generate new types of media that mimic human reasoning in new ways and infiltrate the physical world with a new generation of AI.
In the coming months, AI-powered image generators such as DALL-E and Midjourney will provide video and still images in no time. They will gradually merge with chat AI such as ChatGPT.
This means that chat AI will go far beyond digital text by processing photos, videos, charts, and other media. They will exhibit behaviors more like human reasoning, solving increasingly complex tasks in areas such as math and science. As the technology moves into robotics, it will also help solve problems beyond the digital world.
Many of these advances are already beginning to emerge in top research labs and tech products. But by 2024, these products will be significantly enhanced and used by more people.
"Rapid advances in artificial intelligence. It's going to be here," said David Luan, CEO of Adept, a start-up AI company, "and it's going to be inevitable." ”
OpenAI, Google, and other tech companies are advancing AI. Because of the way the underlying system is built, it's much faster than other technologies.
Most software applications are built by engineers, one line at a time, a process that is often slow and tedious. These tech companies are improving artificial intelligence. Make it faster because the technology relies on neural networks and mathematical systems that can learn skills by analyzing numerical data. By pinpointing patterns in data such as Wikipedia articles, books, and digital text picked from the internet, neural networks can learn to generate text on their own.
This year, these tech companies plan to provide ongoing support for AI. The data in AI systems – including images, sounds, and text – is beyond people's comprehension. As these AIs learn the relationships between various data, they will learn to solve increasingly complex problems and prepare them for life in the physical world.
Still, none of this means that AI will soon be comparable to the human brain. The goal of AI companies and entrepreneurs is to create what they call "artificial general intelligence" – a machine that can do anything the human brain can do – which is clearly a daunting task. Although AI has made rapid progress, it is still in its early stages.
Here's what we can say about how AI is evolving this year. How AI will change this year can be seen from recent developments and many media reports.
Instant video
Until now, AI-powered applications have mostly generated text and static images based on prompts. For example, DALL-E can create photorealistic images in seconds based on requests such as "rhinoceros jumping off the Golden Gate Bridge."
But this year, companies like OpenAI, Google, Meta and New York-based Runway may update image generators to make it possible for people to generate videos as well. These companies have built prototypes of tools that can instantly create videos based on short text prompts.
These tech companies are likely to incorporate the capabilities of image and video generators into chatbots, making chatbots more powerful.
"Multimodal" chatbots
Chatbots and image generators were originally developed as separate tools and are now gradually merging. When OpenAI launched a new version of ChatGPT last year, the chatbot could generate images and text.
Many AI companies are building "multimodal" systems, which means that AI can handle many types of media. These systems learn skills by analyzing photos, text, and potentially other types of media, including graphics, sound, and video, so they can generate their own text, images, and sounds.
And that's not all, since the system is still learning the relationships between different types of media, they will be able to understand one type of media and respond with another. In other words, someone can feed an image into the chatbot and it will respond with text (similar to earlier OCR technology, but with no text content on the input image).
"This technology is going to get smarter and more useful," Ahmad Al-Dahle, head of Meta's generative AI team, "and it's going to do a lot more." ”
Multimodal chatbots can make mistakes, just as text-only chatbots can make mistakes. These AI companies are working to reduce errors and build chatbots that can reason like humans.
Better "reasoning"
When Mr. Altman talks about the leap in artificial intelligence, he is referring to chatbots that are better at "reasoning" because they can take on more complex tasks, such as solving complex mathematical problems and generating detailed computer programs.
The goal is to build systems that solve problems carefully and logically through a series of discrete steps, each building on the next, with a tight logical progression. At least in some cases, that's how humans reason.
Scientists disagree on whether chatbots can actually make such reasoning. Some argue that these systems seem to simply integrate when they repeat behaviors seen in internet data. But OpenAI and other companies are building systems that can more reliably answer complex questions involving subjects such as math, computer programming, physics, and other sciences.
"As systems become more reliable, they will become more popular," said Nick Frosst, a former Google researcher who heads Cohere, a start-up AI company.
If chatbots are better at reasoning, then they can become "AI assistants." ”
Artificial intelligence assistants
When companies study how AI systems can solve complex problems one step at a time, it can also improve the chatbot's ability to use software, apps, and websites instead of you.
Researchers are actually transforming chatbots into a new type of autonomous system called artificial intelligence – AI assistants. This means that chatbots can use software applications, websites, and other online tools, including spreadsheets, online calendars, and travel websites. People can then hand over tedious and repetitive work to the chatbot. But here's the risk: These AI assistants could also take jobs away entirely.
Chatbots already act as assistants on a small scale. They can schedule meetings, edit files, analyze data, and build bar charts. But these tools don't always work as well as they should. When applied to more complex tasks, the AI crashes.
This year, some companies will likely launch more reliable AI assistants. "You should be able to delegate any tedious day-to-day computer work to an AI assistant," says David Luan, CEO of Adept.
This could include tracking expenses in an app like Ledger or keeping track of holidays in an app like a work calendar. In the long run, it will move beyond software and internet services into the field of AI.
Smarter robots
In the past, robots were programmed to perform the same task over and over again, such as picking up boxes that were always the same size and shape. But using the same technology as AI, researchers are giving robots the ability to handle more complex tasks — including those they've never seen before.
Just as chatbots can learn to predict the next word in a sentence by analyzing large amounts of numerical text, robots can learn to predict what will happen in the physical world by analyzing videos of countless objects being stimulated, lifted, and moving.
"These technologies can absorb a lot of data. As they absorb data, they can learn how the world works, how physics works, how they interact with objects. Peter Chen, a former OpenAI researcher and operator of robotics startup Covariant, said.
This year, AI will power robots that operate behind the scenes, such as robotic arms that fold shirts in laundromats or sort piles of items inside a warehouse. Tech giants like Elon Musk are also working hard to bring humanoid robots into people's homes.
#Let's have some dry goods#