Listen to this Post
The Growing Debate Around AI Dependence
Artificial intelligence is becoming more powerful each year, with OpenAI’s GPT-5 now positioned as one of the most advanced language models in existence. Yet, despite its improvements, the company itself is urging caution. Nick Turley, Head of ChatGPT, made it clear in an interview with The Verge that ChatGPT should not replace trusted human expertise. Instead, he insists that it be used as a “second opinion,” because the system, while more refined than ever, still makes errors that could mislead users.
Why ChatGPT Cannot Be Treated as 100% Reliable
OpenAI acknowledges that GPT-5 still struggles with “hallucinations,” where the chatbot generates convincing but factually incorrect information. Although error rates have been reduced, Turley confirmed that incorrect responses still occur in roughly 10% of cases. He stressed that building an AI model that achieves absolute accuracy across every domain is nearly impossible. This is why the company repeatedly emphasizes double-checking AI outputs against reliable human or professional sources.
Steps Taken to Reduce Errors
To tackle misinformation, OpenAI has integrated live search capabilities into ChatGPT. This allows users to cross-check claims with external, real-time information instead of depending solely on the model’s internal knowledge. Turley expressed confidence that hallucinations will eventually be eliminated, but he warned that the solution is not on the immediate horizon. According to him, progress is ongoing, but it may take years before AI reaches human-level reliability in all fields.
Expansion of OpenAI’s Ambitions
Beyond accuracy issues, OpenAI continues to explore new ventures. Reports reveal that the company is developing its own browser, possibly to create a seamless environment where AI and search converge. Adding fuel to speculation, CEO Sam Altman has even hinted that OpenAI would consider acquiring Google Chrome if it ever came up for sale, though this remains more of a bold statement than an immediate plan. These moves indicate OpenAI’s ambition to not only dominate AI but also reshape how people interact with the internet.
Trust Issues and Human Oversight
Large language models like GPT-5 operate by predicting the most likely sequence of words based on vast training datasets. While this method creates natural and human-like conversations, it also increases the risk of producing false claims, particularly on unfamiliar or niche topics. This is precisely why OpenAI is stressing responsible use. AI is a powerful assistant, but it is not a replacement for human judgment.
What Undercode Say:
The statements from Nick Turley highlight a fundamental issue in the AI revolution: trust. GPT-5 is undeniably more sophisticated than its predecessors, but its inability to guarantee full accuracy reinforces why it should be viewed as an aid rather than a replacement. The fact that OpenAI itself advises users not to rely on ChatGPT as their sole source of truth reflects a rare level of corporate transparency, and this honesty could help set realistic expectations for AI adoption.
At its core, the problem lies in how large language models function. They excel at predicting patterns in text but lack the critical reasoning and contextual understanding that human experts bring to the table. This means that even when ChatGPT “sounds” confident, it may still deliver errors that can lead to misinformation if accepted uncritically. Such hallucinations not only damage user trust but also fuel skepticism among professionals in law, medicine, education, and journalism, where factual integrity is non-negotiable.
The integration of real-time search is a vital step forward, as it narrows the gap between AI-generated responses and verified information. However, the reliance on external sources also underscores that ChatGPT alone is insufficient for critical decision-making. Users must become more digitally literate, learning to question AI-generated answers instead of treating them as absolute truths. This shift in mindset will determine whether society benefits from AI responsibly or falls victim to its shortcomings.
Another interesting dimension is OpenAI’s expansion into browsers. By controlling the platform where people access information, the company could streamline fact-checking, blending AI with real-time web data. If successful, this could reduce hallucinations significantly and position OpenAI as a dominant force in both AI and internet navigation. Yet, speculation about buying Chrome is more symbolic than realistic. Google would never casually sell its most powerful digital gateway, but Altman’s remark signals OpenAI’s ambition to challenge established giants.
The timing of these announcements also suggests a strategic narrative. OpenAI is simultaneously managing expectations about GPT-5’s limitations while building hype around its potential ecosystem. By framing ChatGPT as a “second opinion,” the company shifts liability away from itself and empowers users to be cautious. This is a clever communication strategy: it acknowledges weaknesses while reinforcing trust through transparency.
The broader implication is clear. AI, in its current state, cannot replace human expertise. Instead, it should be seen as an augmentation tool — a way to speed up research, explore perspectives, and brainstorm ideas. Professionals who use ChatGPT responsibly can save time and enhance productivity, but those who treat it as infallible risk amplifying mistakes. In other words, AI is not a substitute for wisdom but a partner in knowledge exploration.
From a technological perspective, the claim that hallucinations will not be solved “next quarter” is significant. It signals that OpenAI expects slow, incremental progress rather than breakthroughs. This honesty prevents overhype and prepares users for a long journey toward fully reliable AI. It also opens the door for competitors, as other companies may seek to tackle hallucination problems faster.
In the grand scheme, ChatGPT represents both the promise and peril of artificial intelligence. It has revolutionized the way people engage with information, yet its limitations serve as a reminder that no machine, however advanced, can yet replicate the depth of human judgment. By encouraging a “second opinion” approach, OpenAI is not downplaying its technology’s power; it is guiding users toward responsible adoption. The question now is whether society will listen — or whether the convenience of instant AI answers will outweigh the discipline of fact-checking.
🔍 Fact Checker Results
✅ OpenAI’s GPT-5 is confirmed to still produce around 10% incorrect answers.
✅ Nick Turley publicly advised treating ChatGPT as a “second opinion.”
❌ Rumors about OpenAI buying Chrome are speculative, not confirmed deals.
📊 Prediction
As AI continues to evolve, hallucinations will persist for several more years, though their frequency will decrease steadily. OpenAI’s integration of search and potential browser launch may shift the landscape, making AI a central gateway to online knowledge. However, human expertise will remain irreplaceable, and the future of AI will depend on whether users embrace it as a partner in decision-making rather than a replacement for judgment.
🕵️📝✔️Let’s dive deep and fact‑check.
References:
Reported By: zeenews.india.com
Extra Source Hub:
https://www.reddit.com
Wikipedia
OpenAi & Undercode AI
Image Source:
Unsplash
Undercode AI DI v2
🔐JOIN OUR CYBER WORLD [ CVE News • HackMonitor • UndercodeNews ]
📢 Follow UndercodeNews & Stay Tuned:
𝕏 formerly Twitter 🐦 | @ Threads | 🔗 Linkedin | 🦋BlueSky | 🐘Mastodon
