Sam Altman Claims AI Singularity Arrived, While China's AI Giants Struggle to Profit

Two of today's biggest AI stories pull in opposite directions. On one side, OpenAI CEO Sam Altman declared that AI singularity has arrived, a claim that immediately drew skepticism and debate. On the other, the New York Times reported that even China's leading AI companies are struggling to turn their technology into money.

Altman's Singularity Claim Raises More Questions Than Answers

Sam Altman told ABC News that AI singularity, the point where artificial intelligence begins improving itself without human intervention, is no longer a theoretical concept. It is here.

The claim matters because Altman runs the company behind ChatGPT, the most widely used AI product on the planet. When someone with that kind of influence makes a statement about superintelligence arriving, markets listen. Regulators listen. Competitors listen.

But singularity has a specific technical meaning. It refers to recursive self-improvement, where AI systems rewrite their own code to become more capable, then repeat that cycle without human input. Most researchers say current large language models do not do this. They generate text based on patterns in training data. They do not modify their own architecture or training process.

Altman's framing may be more about narrative than neuroscience. OpenAI is in the middle of a massive funding round that values the company at hundreds of billions of dollars. Declaring singularity achieved is useful when you are selling the story that your product will change everything.

The more grounded takeaway: AI capabilities are advancing fast, but calling current systems "superintelligent" stretches the definition. What Altman may actually be describing is that AI has reached a point where it can do useful work across a wide range of tasks, from writing code to analyzing medical data. That is impressive. It is also not singularity.

China's AI Companies Cannot Crack the Profit Problem

While Altman talks about superintelligence, Chinese AI companies face a more mundane challenge. They cannot make money.

The New York Times reported that firms like Baidu, Alibaba, and ByteDance have built impressive AI models, but none of them have found a way to charge enough to cover the enormous costs of running them. Server bills for AI training and inference run into billions of dollars. Meanwhile, users expect AI features for free or nearly free, because Chinese tech culture has trained them to expect free services subsidised by advertising.

This is the same problem Western AI companies face, but worse. The Chinese market is more competitive, with dozens of companies offering similar AI products. Price wars drive margins to zero. And unlike OpenAI, which can charge $20 per month for ChatGPT Plus, Chinese consumers resist paid AI subscriptions.

The pattern mirrors what happened in ride-hailing, food delivery, and short video. Chinese tech companies pour money into user acquisition, hope their competitors go broke first, and then raise prices. The problem with AI is that the infrastructure costs never go down enough to make this strategy profitable.

For Australian businesses watching from the sidelines, this creates an opportunity. Chinese AI models are likely to remain cheap or free for the foreseeable future because companies are subsidising them to grab market share. That makes tools built on these models affordable for small businesses.

State Attorneys General Take Aim at AI Business Practices

Reuters reported that state attorneys general across the United States are starting to use existing consumer protection and fraud laws to go after AI companies. This is not new AI legislation. They are applying old laws to new problems.

The approach matters because it is faster than passing new rules. A state AG can sue an AI company for deceptive practices under existing consumer protection statutes without waiting for Congress to act. Several states have already sent letters to AI companies demanding transparency about how their products work and what data they collect.

For businesses using AI in the US, this means the legal environment is getting tighter even without new federal legislation. Misleading claims about AI capabilities, hidden data collection, and biased outputs are all potential targets.

AI Adoption Is Outpacing Training in Public Health

A survey by CIDRAP found that field epidemiology programs are adopting AI tools faster than they can train people to use them correctly. Workers are using AI for data analysis, disease surveillance, and report writing, but many programs have not updated their training curricula to cover AI literacy.

This gap creates real risks. An epidemiologist who does not understand how an AI model was trained may misinterpret its output. A public health agency that uses AI to automate report writing without human review could publish inaccurate data.

The pattern repeats across industries. AI tools are easy to deploy. Training people to use them well takes time and money. Most organisations skip the training step and deal with the consequences later.

What This Means for Your Business

The common thread across all four stories is the gap between AI hype and AI reality. Altman sells a vision of superintelligence. Chinese companies cannot make AI profitable. Regulators are using old laws to police new technology. Public health agencies deploy AI without proper training.

None of this means AI is not useful. It is. But businesses that invest in understanding the actual capabilities and limitations of AI tools, rather than chasing the latest headline, will get more value from the technology.

Start with a specific problem. Test AI tools against that problem. Measure the results. Scale what works. That is boring advice, but it is the approach that has worked for every useful technology in the past thirty years.

---

This article was written by Triweb AI's editorial team based on analysis of today's leading AI news sources.