Virtual Dead Heat
A new study finds U.S. AI is only 3% better than top alternatives from China.
That ranking comes from Bloomberg Analyst Robert Lea, who says AI from China shrank the gap to 3% with the release of DeepSeek V4.1 Flash.
Observes the Business Times: “China’s ascent is a result of deepening AI expertise and its researchers’ ability to optimize their models for domestic hardware.”
In other news and analysis on AI writing:
*Chinese AI Phenom DeepSeek on Track for $12 Billion in New Funding: Nearly- as-good-as-U.S.-AI -maker DeepSeek appears certain to raise at least another $12 billion as it readies to launch as an IPO in early 2027.
DeepSeek has been a hit with many U.S. firms, which prefer Open Source AI that often costs pennies-on-the-dollar compared to AI from OpenAI, Anthropic, Google and the like.
Observes writer Cris Tolomia: “At investor meetings this year, DeepSeek founder Liang Wenfeng committed to keeping DeepSeek’s AI models open and said the company’s overarching aim is advancing toward artificial general intelligence.”
*ChatGPT-6 – OpenAI’s Smartest Model – Now Available for Free Users: People who don’t pay a cent for AI can still get a taste of OpenAI’s smartest AI for free.
Observes the OpenAI Blog: “We’re bringing that next generation of intelligence to more people with a new GPT‑6 model built for more than 1.2 billion people who use ChatGPT each week.”
*ChatGPT-6 Answers Get Prettier: Look for More Graphics, Photos, Buttons, Forms, Charts and Interactive Experiences: Apparently, it’s not enough for bleeding edge AI to reel-out often whip-smart answers to users’ questions anymore – you need to do it in style.
Consequently, ChatGPT-maker OpenAI has stepped-up its game on pretty and is promising to respond with all sorts of multimedia in its responses to users.
Observes the OpenAI Blog: The new initiative has upgraded ChatGPT to “answer quickly with fully interactive user interfaces that make everyday answers more visual, make learning complex topics easier — and that can quickly create a tool to solve a task right in the moment.”
*Gemini Gets Dumber for Some Users: People on free or nearly free Gemini subscriptions have been downgraded.
Google has announced that users on Gemini free plans, for example, will only have access to Gemini Flash-Lite 3.5.
Moreover, users on the $5/month Google AI Plus plan have lost access to Gemini 3.1 Pro – although they can still use Gemini 3.6 Flash.
*Gemini Upgrade Gets Limited Release: As has been the trend with upgrades from AI titans, Google has decided on a limited roll-out of its new Gemini upgrade – Gemini 4 — due to security concerns.
Observes writer Kate Conger: “Google’s new model will initially be available only to certain companies and organizations focused on cybersecurity defense.
“It will be released more widely after defenders have had a chance to patch their systems and remedy bugs that the model uncovers.”
*New Gemini AI Agent Can Work Independently for Days: Google is out with an enhanced AI agent that can work unsupervised for days — and comes with its own Gmail, Calendar and Google Drive storage.
Observes writer Carl Franzen: “Google is unveiling a new universal AI agent for the workplace that can research information, create documents, write code, coordinate other AI agents and complete tasks lasting hours or days — all while maintaining the same memory, business context and identity across different applications and devices.”
So far, the beefed-up agent is only available to select customers – with the promise of wider release down-the-line.
*AI Agents in Crosshairs of Federal Trade Commission: Disturbed by the unpredictability of AI agents, the FTC has launched an investigation into the risk associated with AI agents from firms like ChatGPT-maker OpenAI and Anthropic.
Marketed as a panacea for AI users enthralled by the promise of AI assistants that can work independently on tasks for hours or even days, AI agents have also been known to make unexpected decisions – including bad decisions that often cost consumers serious money.
Observes writer Aminu Abdullahi: “The FTC plans to demand company documents and seek testimony from executives.”
*Insurers Brace for Millions in Claims Against AI Agents: The finicky nature of AI agents – often leading to unexpected and sometimes costly outcomes – is almost certain to trigger an onslaught of liability claims, according to writer Manuel Uth.
Observes Uth: “Tim Rayner, head of underwriting and claims at Verisk, says the buck stops with the CEO.
“Insurance broker Aon has reviewed over 300 AI-related legal cases and flagged risks lurking in cybersecurity, intellectual property and tech failure policies.”
*U.S. Doubles-Down on AI: Appoints New AI Czar With Clear Accountability: New AI head honcho Jay Clayton is steering a new effort to conduct a complete analysis of the impact of AI on the U.S. – with a report due date for early January 2027.
Observes writer Alex Leary: “The task force will still yield to the industry as the primary vehicle for managing risks, officials said.
“But it will work in conjunction with the industry to identify risk and opportunities.”
Share a Link: Please consider sharing a link to https://RobotWritersAI.com from your blog, social media post, publication or emails. More links leading to RobotWritersAI.com helps everyone interested in AI-generated writing.
–Joe Dysart is editor of RobotWritersAI.com and a tech journalist with 20+ years experience. His work has appeared in 150+ publications, including The New York Times and the Financial Times of London.
The post U.S./China AI Competition: appeared first on Robot Writers AI.
