AI Agents Promise to Bridge the Gap Between Reality and Science Fiction - Latest Global News

AI Agents Promise to Bridge the Gap Between Reality and Science Fiction

If you have switched on for Google I/O, OpenAI’s Spring Updateor MicrosoftBuild This month, you’ve probably heard the term AI agents quite a bit. They’re quickly becoming the next big thing in technology, but what exactly are they? And why is everyone suddenly talking about them?

Google CEO Sundar Pichai described an artificial intelligence system that return a pair of shoes in your name while on stage at Google I/O. At Microsoft, the company announced copilot AI systems that could act independently like virtual employees. Meanwhile, OpenAI unveiled an AI system, GPT-4 Omni, that can see, hear and speak. Earlier, OpenAI CEO Sam Altman told MIT Technology: helpful agents hold the best potential of technologyThese types of systems are the new benchmark that all AI companies want to achieve, but that’s easier said than done.

Simply put, AI agents are just AI models that do something independently. It’s like Jarvis from Iron ManTar from Interinstaller, or HAL 9000 from A Space Odyssey. They go a step further than just creating a response, like the chatbots we know – there are actions. First, Google, Microsoft and OpenAI are trying to develop agents that can perform digital actions. That is, they teach AI agents to work with various APIs on your computer. Ideally, they can press buttons, make decisions, monitor channels autonomously and send requests.

“I agree that agents are the future,” said Alexander Kvamme, founder and CEO of Echo AI. His company develops AI agents that analyze a company’s conversations with customers and provide insights on how to improve that experience. “The industry has been talking about this for years, but until now it hasn’t happened. It’s just such a difficult problem.”

According to Kvamme, a true agent system must make dozens or hundreds of decisions on its own, which is difficult to automate. For example, to return a pair of shoes, as Google’s Pichai explained, an AI agent might need to search your email for a receipt, get your order number and address, fill out a return form, and perform various actions on your behalf. There are many decisions in this process that you don’t even think about, but make unconsciously.

As we have seen, large language models (LLMs) are not perfect even in controlled environments. Altman’s new favorite pastime is calling ChatGPT “incredibly stupid”, and he’s not entirely wrong. Asking LLMs to work independently on the open internet leaves them vulnerable to error. But that’s exactly what countless startups, including Echo AI, as well as larger companies like Google, OpenAI, and Microsoft are working on.

If you can create agents digitally, there are no major barriers to creating agents that also work with the physical world. You just need to program a robot to do this task. Then you are truly in the realm of science fiction, because AI agents offer the possibility of assigning robots tasks like “take order for this table” or “install all the shingles on this roof.” We are still a long way from that, but the first step is to teach AI agents to perform simple digital tasks.

In the world of AI agents, there is an often discussed problem: making sure that you do not design an agent to do a task, to Well, yes. If you’re building an agent that returns shoes, you need to make sure it doesn’t return all of your shoes, or maybe all of the things you have receipts for in your Gmail inbox. While it may sound silly, there’s a small but vocal group of AI researchers who worry that overly determined AI agents could spell the downfall of human civilization. I suppose when you’re building science fiction, that’s a legitimate concern.

On the other side of the spectrum are optimists like Echo AI who believe this technology will empower us all. This divergence in the AI ​​community is quite large, but the optimists see AI agents as having a liberating effect comparable to that of the personal computer.

“I firmly believe that much of the work that [agents] “We will solve the work that people would rather not do,” said Kvemme. “And they can use their time more wisely. But they have to adapt here too.”

Another use case for AI agents is self-driving cars. Tesla and Waymo are currently the frontrunners of this technology, where cars use AI technology to navigate city streets and highways. Although it is a niche area, self-driving technology is a fairly developed area of ​​AI agents where we are already seeing AI in action in the real world.

So how do we get to the future where AI can give you your shoes back? First, the underlying AI models will likely need to get better and more accurate. This means that updates to ChatGPT, Gemini and Copilot will likely precede fully functional agent systems. AI chatbots still need to expand their huge Hallucination problemfor which many researchers see no solution. But the agent systems themselves also need to be updated. Currently, OpenAI’s GPT store is the most mature attempt to develop an agent network, but even that is not very advanced.

While advanced AI agents are definitely not there yet, this is the goal of many AI companies large and small today. This could make AI significantly more useful in our everyday lives. Although it sounds like science fiction, billions of dollars are being spent to make agents a reality in our lifetimes. However, this is a tall order for AI companies that have struggled to get chatbots to reliably answer basic questions.

Sharing Is Caring:

Leave a Comment