Jensen Huang argues American companies should be allowed to use Chinese AI models — Nvidia CEO says backdoors connected to China are misconceptions ⁠71%

By Jowi Morales⁠76%

7/22/2026, 5:55:46 PM

BS Summary: This article contains 25 faulty reasoning types, including Post Hoc (False Cause), Negativity Bias, and Optimism Bias, with Anecdotal as the most egregious example at 23.1% saturation with 115 hits. Analysis detected 949 faulty-reasoning hits from 498 analyzed words, generating a BS Score of 64.3% and a BS Rank of ⁠71% (5,960 of 20,518 articles). This article is worse (more manipulative) than 71.00% of the article peer group.

Nvidia CEO Jensen Huang thinks that American companies should be allowed to use Chinese AI models, even as Washington is trying to ban them . 
When Axios co-founder Mike Allen asked Huang in an interview if Americans companies should be allowed to use Chinese AI models, Huang responded with “absolutely.” 
The answer comes right after Chinese firm Moonshot AI released a 2.8T open-weight model called Kimi K3 , which — although it isn’t as powerful as frontier models like Fable 5 — is comparable to GPT 5.5 and Claude Opus 4.8 while costing just a third of these models. 
One of the biggest concerns of U.S. leaders have is that these AI models might come with vulnerabilities that the Chinese government can use to attack American interests, but Huang said that this is an incorrect assumption. 
“There is a misconception that somehow there are backdoors that are somehow connected to China in some way,” said the Nvidia chief. 
“You download the models, you can fine-tune it, you can enhance it, you can guardrail it as you desire.” 
Huang shares the same sentiments about American AI models. 
Just last month, the U.S. enforced an export restriction on Anthropic’s Mythos and Fable 5 , citing security threats — although access was eventually restored after its developer placed a filter to block these tools from identifying software vulnerabilities. 
OpenAI’s ChatGPT-5.6 received the same treatment, and Washington warned the firm that it should not release its latest model without getting the green light from the government. 
Huant argues that, instead of restricting access to these powerful models at launch, AI firms should make their models available to all and make them more secure through rapid testing and fixes. 
But even as he advocated the need for everyone to have access to closed models, Jensen also noted that various industries, such as the sciences and cybersecurity, need open models as well. 
He claims that these models make AI more secure, as other people can inspect them to look for weaknesses and fix them as required. 
“If everything just becomes one single model, one single point of attack, one single source of failure, I think the world is much, much more vulnerable,” Huang said. 
As for the market’s negative reaction every time cheaper, open-weight models become available, the Nvidia CEO says that investors misunderstand their impact. 
Huang said that this happened when DeepSeek arrived for the first time, and it’s happening again with the arrival of Kimi. 
He says that these open models, which cost less to run, will encourage more people to use AI. 
So instead of cutting data center demand, these cheaper, more efficient models are actually good for the industry in general because they will drive demand. 
(And with higher demand, there’s more incentive to build data centers and buy AI GPUs, which is ultimately good for Nvidia.) 
Article reasoning-pattern comparisonThis article: 4.4%Jowi Morales: 4.1%Tom's Hardware: 3.5%Confirmation Bias4.4%This article: 0.0%Jowi Morales: 2.6%Tom's Hardware: 2.3%Anchoring Bias0.0%This article: 5.6%Jowi Morales: 4.8%Tom's Hardware: 3.4%Availability Heuristic5.6%This article: 9.8%Jowi Morales: 1.4%Tom's Hardware: 1.1%Representativeness Heuristic9.8%This article: 0.0%Jowi Morales: 0.5%Tom's Hardware: 0.5%Hindsight Bias0.0%This article: 8.6%Jowi Morales: 3.4%Tom's Hardware: 3.5%Overconfidence Bias8.6%This article: 4.4%Jowi Morales: 13.1%Tom's Hardware: 9.5%Framing Effect4.4%This article: 0.0%Jowi Morales: 2.6%Tom's Hardware: 1.0%Loss Aversion0.0%This article: 6.4%Jowi Morales: 1.0%Tom's Hardware: 0.7%Status Quo Bias6.4%This article: 0.0%Jowi Morales: 0.1%Tom's Hardware: 0.3%Sunk Cost Effect0.0%This article: 10.0%Jowi Morales: 7.1%Tom's Hardware: 5.8%Optimism Bias10.0%This article: 0.0%Jowi Morales: 2.3%Tom's Hardware: 1.8%Pessimism Bias0.0%This article: 12.2%Jowi Morales: 14.0%Tom's Hardware: 6.5%Negativity Bias12.2%This article: 9.2%Jowi Morales: 1.4%Tom's Hardware: 1.3%Self-Serving Bias9.2%This article: 0.0%Jowi Morales: 0.9%Tom's Hardware: 0.6%Fundamental Attribution Error0.0%This article: 0.0%Jowi Morales: 0.2%Tom's Hardware: 0.2%Actor-Observer Bias0.0%This article: 0.0%Jowi Morales: 1.3%Tom's Hardware: 0.5%In-Group Bias0.0%This article: 0.0%Jowi Morales: 0.3%Tom's Hardware: 0.2%Out-Group Homogeneity Bias0.0%This article: 9.8%Jowi Morales: 2.8%Tom's Hardware: 3.0%Halo Effect9.8%This article: 0.0%Jowi Morales: 0.0%Tom's Hardware: 0.0%Horn Effect0.0%This article: 0.0%Jowi Morales: 0.0%Tom's Hardware: 0.0%Dunning-Kruger Effect0.0%This article: 4.2%Jowi Morales: 1.8%Tom's Hardware: 2.0%Recency Bias4.2%This article: 0.0%Jowi Morales: 0.7%Tom's Hardware: 0.6%Primacy Effect0.0%This article: 0.0%Jowi Morales: 0.0%Tom's Hardware: 0.1%Blind-Spot Bias0.0%This article: 0.0%Jowi Morales: 0.3%Tom's Hardware: 0.2%Ad Hominem0.0%This article: 7.4%Jowi Morales: 0.5%Tom's Hardware: 0.2%Straw Man7.4%This article: 0.0%Jowi Morales: 6.8%Tom's Hardware: 5.1%Appeal to Authority0.0%This article: 6.4%Jowi Morales: 4.0%Tom's Hardware: 2.0%False Dilemma6.4%This article: 0.0%Jowi Morales: 1.5%Tom's Hardware: 0.9%Slippery Slope0.0%This article: 4.8%Jowi Morales: 0.2%Tom's Hardware: 0.1%Circular Reasoning4.8%This article: 7.4%Jowi Morales: 9.0%Tom's Hardware: 5.8%Hasty Generalization7.4%This article: 0.0%Jowi Morales: 0.3%Tom's Hardware: 0.3%Red Herring0.0%This article: 5.0%Jowi Morales: 1.1%Tom's Hardware: 1.0%Bandwagon5.0%This article: 0.0%Jowi Morales: 7.4%Tom's Hardware: 3.1%Appeal to Emotion0.0%This article: 4.4%Jowi Morales: 0.8%Tom's Hardware: 0.8%Begging the Question4.4%This article: 16.3%Jowi Morales: 6.0%Tom's Hardware: 3.6%Post Hoc (False Cause)16.3%This article: 0.0%Jowi Morales: 0.3%Tom's Hardware: 0.1%Tu Quoque0.0%This article: 0.0%Jowi Morales: 1.5%Tom's Hardware: 0.7%Burden of Proof0.0%This article: 0.0%Jowi Morales: 0.5%Tom's Hardware: 0.3%Appeal to Nature0.0%This article: 5.6%Jowi Morales: 0.7%Tom's Hardware: 0.6%Composition/Division5.6%This article: 23.1%Jowi Morales: 3.8%Tom's Hardware: 1.7%Anecdotal23.1%This article: 0.0%Jowi Morales: 0.0%Tom's Hardware: 0.0%No True Scotsman0.0%This article: 4.4%Jowi Morales: 3.7%Tom's Hardware: 3.4%Ambiguity (Equivocation)4.4%This article: 0.0%Jowi Morales: 0.0%Tom's Hardware: 0.0%Gambler’s Fallacy0.0%This article: 6.4%Jowi Morales: 0.1%Tom's Hardware: 0.2%Middle Ground6.4%This article: 0.0%Jowi Morales: 0.0%Tom's Hardware: 0.0%Personal Incredulity0.0%This article: 0.0%Jowi Morales: 0.4%Tom's Hardware: 0.2%Special Pleading0.0%This article: 0.0%Jowi Morales: 0.2%Tom's Hardware: 0.1%Genetic Fallacy0.0%This article: 5.0%Jowi Morales: 3.5%Tom's Hardware: 2.1%Unattributed Quote5.0%This article: 5.0%Jowi Morales: 2.7%Tom's Hardware: 1.0%Quote-first Misdirection5.0%This article: 0.0%Jowi Morales: 9.1%Tom's Hardware: 6.9%Biased Writer Voice0.0%This article: 0.0%Jowi Morales: 2.3%Tom's Hardware: 1.4%Indoctrination0.0%This article: 0.0%Jowi Morales: 0.5%Tom's Hardware: 0.1%Politically Left Leaning Bias0.0%This article: 0.0%Jowi Morales: 0.0%Tom's Hardware: 0.0%Politically Right Leaning Bias0.0%This article: 4.2%Jowi Morales: 7.4%Tom's Hardware: 3.4%Attempt to Sell a Product or S…4.2%

498 words analyzed.

Speakers

2speakers54%attributed speech227writer words
Voice mapSelect a segment to jump to its words
Writer's voice • 23 words • 0.0% coverageJensen Huang • 25 words • 0.0% coverageMike Allen • 25 words • 100.0% coverageWriter's voice • 49 words • 0.0% coverageWriter's voice • 37 words • 0.0% coverageJensen Huang • 22 words • 0.0% coverageJensen Huang • 19 words • 0.0% coverageWriter's voice • 9 words • 0.0% coverageWriter's voice • 39 words • 0.0% coverageWriter's voice • 27 words • 0.0% coverageJensen Huang • 32 words • 0.0% coverageJensen Huang • 32 words • 0.0% coverageJensen Huang • 24 words • 0.0% coverageJensen Huang • 28 words • 0.0% coverageWriter's voice • 22 words • 0.0% coverageJensen Huang • 21 words • 0.0% coverageJensen Huang • 18 words • 0.0% coverageJensen Huang • 25 words • 0.0% coverageWriter's voice • 21 words • 100.0% coverage
Selected voice

Mike Allen

100%flagged-word coverage
25 attributed words9.2% of attributed speech86% writer coverage
0%50.0%100.0%Unattributed Quote+100.0 ptsWriter: 0.0%Mike Allen: 100.0%100.0%Quote-first Misdirection+100.0 ptsWriter: 0.0%Mike Allen: 100.0%100.0%Attempt to Sell a Product -9.3 ptsWriter: 9.3%Mike Allen: 0.0%0.0%

Attribution is sentence-level. Pattern percentages are calculated only from words assigned to that voice.

Loading…
Loading…
Loading…

Analysis

Hover over highlighted words in the article to view the associated bias or fallacy analysis.