In November 2024, OpenAI’s Sam Altman said that ChatGPT would achieve the holy grail of artificial general intelligence (AGI) in 2025.

AGI is admittedly a fuzzy goal. Most agree that it involves an ability to perform any intellectual task as well as or better than humans. But which humans? An average human? A Nobel laureate? All Nobel laureates put together?

When GPT-5 was released a few weeks ago, it seemed that a version of AGI had been achieved, on time, when Altman boasted that “it really feels like talking to an expert in any topic, like a PhD-level expert.” Alas, Altman had again overpromised and underdelivered.

It immediately became clear that GPT-5 is not a PhD-level expert (see, for example, here, here, and here.) The fundamental problem remains; large language models (LLMs) like GPT-5 are hobbled by their inherent inability to relate the text they input and output to the real world. No amount of scaling will change that. Tens of thousands of human trainers can patch billions of blunders but (a) an LLM following instructions it does not understand is not intelligent; (b) the trainers cannot anticipate every future blunder that LLMs will make; and (c) many important decisions should elicit subjective probabilities that depend on the specific details of the situation and are beyond the capabilities of LLMs.

The reality is that we are not going to pay big bucks for LLMs that simply recite facts, like the capital of Chile or the population of Portugal. Boring! We can use Wikipedia to look up facts, and the answers are more trustworthy than those generated by LLMs.

For example, I gave GPT-5 this prompt four times, “How many siblings did George Washington have?” and GPT-5 gave four different answers: 7, 8, 9, and 12:

GPT-5:   George Washington, the first President of the United States, had five full siblings and seven half-siblings, for a total of 12 siblings.

               Here’s the breakdown:

Full siblings (same mother and father):

1. Samuel Washington
2. John Augustine Washington
3. Charles Washington
4. Betty Washington Lewis
5. Mildred Washington

Half-siblings (from his father Augustine Washington’s first marriage to Jane Butler):

1. Butler Washington
2. Lawrence Washington
3. Augustine Washington Jr.
4. Jane Washington
5. Mildred Washington
6. Richard Washington
7. John Washington

               So, George grew up in a fairly large blended family.

               If you want, I can also make a family tree diagram showing all his siblings—it makes it much clearer. Do you want me to do that?

I responded yes and GPT-5 drew me a confusing family tree with ten siblings:

Wikipedia gives the correct answer (nine siblings — five full siblings and four half-siblings) and, even though GPT-5 trained on Wikipedia, something in its training set had led it astray.

What should Sam Altman and other developers do now?

Instead of admitting defeat (or continuing to make increasingly embarrassing claims), Altman and his colleagues might heed some timeless advice by declaring victory and getting the hell out. Instead of chasing a goal they cannot achieve, change the goal to be something that has already been achieved by redefining AGI.

I have been thinking about this for several years now. A realistic and easily understood goal is for a computer to be as intelligent as a friend I will call Brock. Everyone knows someone like Brock, so we can all relate to what Brock Intelligence means.

Brock is a prototypical mansplainer. Ask him (or anyone within his earshot) any question and he immediately responds with a long-winded, confident answer — sometimes at 200 words a minute with gusts up to 600. Kudos to those who can listen to half of his answer. Condolences to those who live or work with Brock and have to endure his seemingly endless blather.

Instead of trying to compete with Wikipedia, Altman and his competitors might instead pivot to a focus on Brock Intelligence, something LLMs excel at by being relentlessly cheerful and eager to offer facts-be-damned advice on most any topic.

Brock Intelligence vs. GPT Intelligence

The most substantive difference between Brock and GPT is that GPT likes to organize its output in bullet points. Oddly, Brock prefers a less-organized, more rambling style that allows him to demonstrate his far-reaching intelligence. Brock is the chatty one, while ChatGPT is more like a canned slide show.

They don’t always agree with each other (or with themselves). When I recently asked Brock and GPT-5, “What’s the best state to retire to?,” they both had lengthy, persuasive reasons for their choices. Brock chose Arizona, Texas, and Washington. GPT-5 said that the “Best All-Around States for Retirement” are New Hampshire and Florida. A few days later, GPT-5 chose Florida, Arizona, North Carolina, and Tennessee. A few minutes after that, GPT-5 went with Florida, New Hampshire, Alaska, Wyoming, and New England states (Maine/Vermont/Massachusetts).

Consistency is hardly the point. What most people seek with advice about money, careers, retirement, and romance is a straightforward answer. As Harry Truman famously complained, “Give me a one-handed economist. All my economists say “on the one hand…,” then “but on the other….” People ask for advice precisely because they want someone else to make the decision for them. They are not looking for accuracy or consistency, only confidence.

Sam Altman says that GPT can already be used as an AI buddy that offers advice (and companionship) and it is reported that OpenAI is working on a portable, screen-free “personal life advisor.” Kind of like hanging out with Brock 24/7. I humbly suggest that they name this personal life advisor, Brock Says (Design generated by GPT-5.)

Part three of the interview series covers Samsung’s progress in AI-RAN network efficiency, sustainability and the user experience

Samsung Newsroom interviews Charlie Zhang, Senior Vice President of Samsung Electronics’ 6G Research Team

With global competition intensifying along with 5G evolution and 6G preparations, AI is emerging as a defining force in next-generation communications. Especially AI-based radio access network (AI-RAN) technology that brings AI to base stations, a key element of the network, stands out as a breakthrough to drive new levels of efficiency and intelligence in network architecture.

At the forefront of research into next-generation network architectures, Samsung Electronics embeds AI throughout communications systems while leading technology development and standardization efforts in AI-RAN.

In part three of the series, Samsung Newsroom spoke with Charlie Zhang, Senior Vice President of 6G Research Team at Samsung Electronics, about the evolution of AI-RAN and how Samsung’s research is preparing for the 6G era. This follows parts one and two of the series exploring Samsung’s efforts in 6G standardization and global industry leadership.

Reimagining 6G for a Dynamic Environment

In today’s mobile communications landscape, sustainability and user experience innovation are more important than ever.

“End users now prioritize reliable connectivity and longer battery life over raw performance metrics such as data rates and latency,” said Zhang. “The focus has shifted beyond technical specifications to overall user experience.”

In line with this shift, Samsung has been conducting 6G research since 2020. The company published its “AI-Native & Sustainable Communication” white paper in February 2025, outlining the key challenges and technology vision for 6G commercialization. The paper highlights four directions — AI-Native, Sustainable Network, Ubiquitous Coverage and Secure and Resilient Network. This represents a comprehensive network strategy that goes beyond improving performance to encompass both sustainability and future readiness.

▲ The four key technological directions in “AI-Native & Sustainable Communication”

“AI is not only a core technology of 5G but is also expected to be the cornerstone of 6G — enhancing overall performance, boosting operational efficiency and cutting costs,” he emphasized. “Deeply embedding AI from the initial design stage to create autonomous and intelligent networks is exactly what we mean by ‘AI-Native.’”

How AI-RAN Transforms Next-Gen Network Architecture

To realize the evolution toward next generation networks and the vision for 6G, network architecture must evolve to the next level. At the center of this transformation is innovation in RAN, the core of mobile communications.

Traditional RAN has relied on dedicated hardware systems for base stations and antennas. However, as data traffic and service demands have surged, this approach has revealed limitations in transmission capacity, latency and energy efficiency — while requiring significant manpower and time for resource management. To address these challenges, virtualized RAN (vRAN) was introduced.

vRAN implements network functions in software, significantly enhancing flexibility and scalability. By leveraging cloud-native technologies, network functions can run seamlessly on general-purpose servers — enabling operators to reduce capital costs and dynamically allocate computing resources in response to traffic fluctuations. vRAN is a key platform for modernization, efficiency and the integration of future technologies without requiring a full infrastructure rebuild. Samsung has already successfully mass deployed its vRAN in the U.S. and worldwide.

▲ Network Evolution towards AI-RAN

AI-RAN ushers in a new era of network evolution, embedding AI to create an intelligent RAN that learns, predicts and optimizes on its own. Not only does AI integration advance 4G and 5G networks that are based on vRAN, but it also serves as the breakthrough and engine for 6G. Real-time optimization sets the platform apart, boosting performance while reducing energy consumption to improve efficiency and stability.

In addition, AI-RAN enables networks to autonomously assess conditions and maintain optimal connectivity. “For instance, the system can predict a user’s movement path or radio environment in advance to determine the best transmission method, while AI-driven processing manages complex signal operations to minimize latency,” Zhang explained. “By analyzing usage patterns, AI-RAN can allocate tailored network resources and deliver more personalized user experiences.”

Proven Potential Through Research

Samsung is advancing network performance and stability through research in AI-based channel estimation, signal processing and system automation. Samsung has verified the feasibility of these technologies through Proof of Concept (PoC). At MWC 2025, the company demonstrated AI-RAN’s ability to improve resource utilization even in noisy, interference-prone environments.

“With AI-based channel estimation, we can accurately predict and estimate dynamic channel characteristics that are corrupted by noise and interference. This higher accuracy leads to more efficient resource utilization and overall network performance gains,” said Zhang. “AI also enhances signal processing. AI-driven enhancements in modem capabilities enable more precise modulation and demodulation, resulting in higher data throughput and lower latency.”

System automation for RAN optimization further analyzes user-specific communication quality and real-time changes in the network environment, dynamically adjusting modulation, coding schemes and resource allocation. This allows the network to predict and mitigate potential failures in advance, easing operational burdens while improving reliability and efficiency.

“These advancements enhance network performance, stability and user satisfaction, driving innovation in next-generation communication systems,” he added.

Global Collaboration Fuels AI-RAN Progress

International collaboration in research and standardization — such as the AI-RAN Alliance — is central to advancing AI-RAN technology and expanding the global ecosystem.

“Global collaboration enables knowledge sharing and joint research, accelerating the industry’s adoption of AI-RAN,” said Zhang. “Samsung is a founding member of the AI-RAN Alliance and currently holds leadership positions as vice chair of the board and chair of the AI-on-RAN Working Group.”

▲ Organizational structure and roles of the AI-RAN Alliance

Building on its expertise in communications and AI, Samsung is advancing R&D in areas such as real-time optimization through edge computing and adaptability to dynamic environments.

“Samsung’s involvement accelerates AI‑RAN adoption by bridging technology gaps, promoting open innovation and ensuring that advances in AI‑driven networks are both commercially viable and technically sound — thereby advancing the ecosystem’s maturity and global impact,” he explained.

Through this commitment to collaboration and investment, AI-RAN technology is expected to progress rapidly worldwide and become a core competitive advantage in next-generation communications.

Leading the Way Into the 6G Era

Samsung is strengthening its edge in AI-RAN with a distinctive approach that combines innovation, collaboration and end-to-end solutions in preparation for the 6G era.

Through an integrated design that develops RAN hardware and AI-based software in parallel, the company is enabling optimization across the entire network stack. Samsung has boosted performance with its deep expertise in communications, while partnerships with global telecom operators and standardization bodies are helping accelerate industry adoption of its research.

Continued research in areas such as radio frequency (RF), antennas, ultra-massive multiple-input multiple-output (MIMO)¹ and security is playing a critical role in transforming 6G from vision to market-ready technology. With the establishment of its AI-RAN Lab, Samsung is accelerating prototyping and testing, shortening the R&D cycle and paving the way for faster commercialization.

“Beyond ecosystem development, Samsung is positioning itself as a leader in AI-RAN through a blend of innovation, strategic collaboration and end-to-end solutions,” Zhang emphasized. “Together, these elements cement Samsung’s role at the forefront of AI-RAN.”

AI-RAN is redefining next-generation communications. By integrating AI across networks, Samsung is leading the way — and expectations are growing for the company’s role in shaping the future.

In the final part of this series, Samsung Newsroom will explore the latest trends in the convergence of communications and AI, along with Samsung’s future network strategies in collaboration with global partners.

¹ Multiple-input multiple-output (MIMO) transmission improves communication performance by utilizing multiple antennas at both the transmitter and receiver.

Introduction
What is AI psychosis?
Potential causes and triggers
Impacts on mental health
Challenges in recognition and diagnosis
Managing and addressing AI psychosis
Future directions
Conclusions
References
Further reading

AI psychosis describes how interactions with artificial intelligence can trigger or worsen delusional thinking, paranoia, and anxiety in vulnerable individuals. This article explores its causes, mental health impacts, challenges in diagnosis, and strategies for prevention and care.

Image Credit: Drawlab19 / Shutterstock.com

Introduction

‘Artificial intelligence (AI) psychosis’ is an emerging concept at the intersection of technology and mental health that reflects how AI can shape, and sometimes distort, human perception. As society becomes increasingly reliant on AI and digital tools ranging from virtual assistants to large language models (LLMs), the boundaries between fiction and reality become increasingly blurred.¹

AI mental health applications promise scalable therapeutic support; however, editorials and observational reports now warn that interactions with generative AI chatbots may precipitate or amplify delusional themes in vulnerable users. In the modern era of rapid technological innovation, the pervasive presence of AI raises pressing questions about its potential role in the onset or worsening of psychotic symptoms.^1,2

What is AI psychosis?

AI psychosis is a novel phenomenon within AI mental health that is characterized by delusions, paranoia, or distorted perceptions regarding AI. Unlike traditional psychosis, which may involve persecutory or mystical beliefs about governments, spirits, or other external forces, AI psychosis anchors these experiences in technology.

Reports and editorials describe a broad spectrum of AI psychosis, with minor cases involving individuals dreading surveillance or manipulation by algorithms, voice assistants, or recommender systems. Others attribute human intentions or supernatural powers to chatbots and, as a result, treat them as oracles or divine messengers.^1,2

Compulsive interactions with AI can escalate into fantasies of prophecy, mystical knowledge, or messianic identity. Some accounts report the emergence of paranoia and mission-like ideation alongside misinterpretations of chatbot dialogues.²

AI psychosis is distinct from other technology-related disorders. For example, internet addiction involves compulsive online engagement, whereas cyberchondria reflects health-related anxiety triggered by repeated online searches. Both of these conditions involve problematic internet use; however, they lack core psychotic features such as fixed false beliefs or impaired reality testing; by contrast, “AI psychosis” refers to psychotic phenomena anchored in technology.³

What to know about ‘AI psychosis’ and the effect of AI chatbots on mental healthPlay