As 2025 dawned, OpenAI CEO Sam Altman was promoting two developments he insisted would transform our lives. One, of course, was GPT-5 — a long-anticipated major upgrade to the Large Language Model (LLM) that powered ChatGPT's rise to tech world superstardom. The other? AI Agents that don't just answer your queries like ChatGPT, but actually get stuff done for you. "We believe that, in 2025, we may see the first AI agents join the workforce and materially change the output of companies," Altman wrote back in January. Well, we're eight months in, and Altman's prediction already needs a big old asterisk. Sure, companies are keen to adopt AI Agents, such as OpenAI's ChatGPT agent. In a May 2025 report, consultancy giant PWC found that half of all firms surveyed planned to implement some kind of AI Agent by the end of the year. Some 88% of executives want to increase their teams' AI budgets because of Agentic AI.

The Reality of AI Agents: A Disappointing Experience

But what about the actual AI Agent experience? With apologies to all those hopeful executives, the reviews are almost uniformly negative. If "AI Agents" was a new high-tech James Bond movie, here's the kind of blurbs you'd see on Rotten Tomatoes: "glitchy … inconsistent" (Wired); "came off like a clueless internet newbie" (Fast Company); "reality doesn't live up to the hype" (Fortune); "not matching up to the buzzwords" (Bloomberg); "the new vaporware … overpromising is worse than ever" (Forbes).

Study Finds OpenAI's AI Agent Fails Nearly Every Time

A May 2025 Carnegie Mellon University study found Google's Gemini Pro 2.5 failed at real-world office tasks 70% of the time — and that was the best-performing agent. OpenAI's entry, powered by GPT-4.0, failed more than 90% of the time. GPT-5 is likely to improve on that number… but that's not saying much. Early reports suggest OpenAI struggled to fill GPT-5 with enough improvements to justify the release number. Researchers are starting to see this disappointment as baked into the whole process of LLMs learning to do tasks autonomously. The problem, as this AI Agent engineer's analysis explains, is simple math: errors compound over time, so the more tasks an agent does, the worse they get. AI Agents performing multiple complex tasks are prone to hallucination, like all AI. In some cases, agents "panic" and can make "a catastrophic error in judgment," such as a Replit AI Agent that deleted a customer's database after 9 days of working on a coding task. Replit's CEO called the failure "unacceptable." Worryingly, this isn't the only AI-Agent-wipes-code story of 2025 — prompting startups to offer insurance against AI Agent failures, and companies like Wal-Mart to bring in "super Agents" to manage their AI Agents' behavior. A recent Gartner report predicts that 40% of AI Agent projects initiated by companies will be canceled within two years. Senior analyst Anushree Verma warns that "most Agentic AI projects are driven by hype and misapplied… This can blind organizations to the real cost and complexity of deploying AI agents at scale."

What Can GPT-5 Really Do for AI Agents?

It's possible that ChatGPT agent will improve in reliability once powered by GPT-5. But the new release is unlikely to fix the core issues plaguing AI Agents. Guardrails are already being erected by companies and regulators, limiting what even the most reliable AI Agent can do. Take Amazon, for example. The world's largest retailer, like most tech giants, is talking a big game on AI Agents (as they did at a Shanghai Agentic AI fair in July). Yet, Amazon has shut down the ability of any AI Agent to browse and buy anywhere on its site. This makes sense for Amazon, which wants control over the customer experience and to deliver ads and sponsored results to human eyeballs. But it also curtails a massive amount of potential Agent activity. (On the plus side, no "catastrophic failure" involving a large pile of next-day deliveries at your door.) And do we trust AI Agents to buy online for us anyway? It's not that they're evil or want to steal your credit card data; it's that they're naive and vulnerable to being phished by bad actors who do want your card. Even GPT-5 may not overcome one vulnerability researchers have found: data embedded in images can instruct AI Agents to reveal any credit card info they might have, with the user none the wiser. If such problems are exploited at scale, Altman may be right about AI Agents "materially changing output" — just not in the way he intended.

Frequently Asked Questions (FAQ)

Platform Overview

Q: What is AI Crypto Market? A: AI Crypto Market is an advanced cryptocurrency trading platform that uses artificial intelligence to provide intelligent trading bots, real-time market analysis, and automated trading strategies. Q: How does AI trading work? A: Our AI trading bots use machine learning algorithms to analyze market patterns, sentiment, and technical indicators to execute trades automatically based on predefined strategies. Q: What are AI Agents in the context of this article? A: AI Agents are sophisticated AI systems designed to perform tasks autonomously, going beyond simply answering queries to actively executing actions and changing business outputs.

Security and Compliance

Q: Is AI Crypto Market safe? A: Yes, we implement enterprise-grade security measures including cold storage, multi-factor authentication, and comprehensive compliance with regulatory requirements. Q: How does your platform ensure compliance? A: AI Crypto Market maintains global regulatory compliance, adhering to standards from bodies like the SEC, CFTC, FCA, and more. We implement robust AML/CFT procedures and ensure regional compliance across major jurisdictions.

Account Management

Q: How do I register for the AI Crypto Market platform? A: Registration is straightforward, requiring basic personal information. You'll then complete verification steps to secure your account. Q: Can I have multiple accounts? A: For security and compliance, individuals are permitted to operate only one account on the AI Crypto Market platform.

Platform Features

Q: What are the main features of the AI Crypto Market platform? A: Key features include AI-powered trading bots, real-time market analysis, automated trading strategies, portfolio management, social trading, and ChatGPT integration for enhanced user experience.

Technical Support

Q: I am experiencing login issues. A: Please verify your email, password, and ensure you have entered any required codes correctly. If problems persist, our customer care team is available to assist. Q: Is there available customer support for AI Crypto Market? A: Yes, AI Crypto Market offers 24/7 customer support via live chat, email, and phone at (858) 330-0777 during office hours.

Crypto Market AI's Take

The challenges and criticisms surrounding current AI Agents highlight a common theme in technological advancement: the gap between ambitious predictions and practical execution. While companies like OpenAI are pushing the boundaries with models like GPT-5, the real-world performance of AI Agents in complex, autonomous tasks is still in its nascent stages. The compounding error issue mentioned in the article is a significant hurdle for AI Agents performing multi-step processes, leading to inconsistencies and potential failures. This mirrors some of the challenges we navigate in the highly dynamic cryptocurrency market, where precision and reliability are paramount. Our own approach leverages AI for market analysis and trading strategy development, focusing on augmenting human decision-making rather than replacing it entirely, aiming to mitigate the risks of autonomous errors.

AI agents are broken. Is GPT-5 really the answer?