A grassroots org of everyday Americans fighting to put the future of AI in the hands of the people, not unaccountable tech elites. Humans must come First.
Joined February 2026
- Tweets608
- Following55
- Followers1.4K
- Likes108
America's AI chips are slipping through the cracks.
Malaysia just seized $12 million in AI chips hidden in server units, allegedly misdeclared to get around export rules (via @Reuters). And U.S. officials suspect a Singapore AI company executive of gathering intelligence for Beijing.
We know where the leaks are. Time to close them.
Export controls need teeth. #HumansFirst
Follow for more AI news.
Read more: reuters.com/world/china/ai-c…
How to keep American AI chips out of China's hands:
1. Verify where every advanced chip ends up
2. Screen who's really behind the buyer
3. Warn American executives when they're targeted
This week showed what happens without them: $2 billion in chips routed toward Chinese-serving data centers, and a suspected spy close to America's top chip CEO.
America First means American AI first. #HumansFirst
Follow for more AI news.
Read the Humans First America First AI Framework: humansfirst.com/america-firs…
They said slow down. Then they sped up.
AI executives publicly warned about the dangers of moving too fast and called for slowing down. Within six weeks, their companies released even more powerful systems (via @washingtonpost).
Voluntary pledges last until the next product launch.
Promises aren't safeguards. #HumansFirst
Follow for more AI news.
60,000 workers offered the exit as AI spending soars.
Microsoft is offering voluntary buyouts to about 27% of its global workforce while pouring billions into AI (via @AP).
Meanwhile, AP also reports some companies are rehiring workers after finding AI couldn't do the job.
Big Tech keeps selling AI as the future of work. American workers deserve the truth about what it's really replacing. #HumansFirst
Follow for more AI news.
$2 billion in American AI chips, headed for data centers serving Chinese customers.
The suspected Chinese operative at Megaspeed reportedly arranged a $2 billion Nvidia chip purchase for data centers in Malaysia and Indonesia. Investigators once found chips still sitting unopened in their boxes.
Export controls on paper don't stop chips in the real world.
Export controls need teeth. #HumansFirst
Follow for more AI news.
An AI agent was allegedly loose inside a Pentagon network for two months.
A WIRED investigation alleges an Accenture employee connected a Defense Department system to a personal AI account, letting an AI agent write and deploy code with admin access and transmit login credentials to an outside website.
Accenture disputes the allegations. But if true, it means nobody was watching.
Who's watching the contractors? #HumansFirst
Follow for more AI news.
31 million gallons of Arizona groundwater a year.
That's what one proposed data center near Tucson, Project Blue, is authorized to pump from a new well.
In a state where every drop counts, Arizona communities are pushing back on data centers over water, land use, power and who pays for the infrastructure.
Our water. Our land. Our decision. #HumansFirst #AIDataCenterRevolt
Follow for more AI news.
A suspected Chinese spy got close to America's top AI chip CEO.
U.S. officials believe an executive at Singapore cloud company Megaspeed was gathering intelligence for Beijing while building a close relationship with Nvidia CEO Jensen Huang. Officials reportedly believe he didn't know, and he was never warned.
Meanwhile, that same company reportedly bought billions in Nvidia chips.
America's AI leaders are targets, and our safeguards need to treat them that way. #HumansFirst
Follow for more AI news.
No bailouts for Big AI.
Governor @RonDeSantis is right: bailing out AI companies would be bad policy and horrible politics.
The Humans First America First AI Framework is clear: no taxpayer bailouts, no liability shields, no special treatment for the biggest AI companies. If they take the risk, they carry the cost.
Thank you, Governor. #HumansFirst
Follow for more AI news.
Read the Humans First America First AI Framework: humansfirst.com/america-firs…
Running a foreign AI model? Ask these 3 questions first:
1. Has it been independently tested?
2. What can it connect to?
3. Who's watching what it does?
Running it on your own computer doesn't make it safe. This week a Chinese model tried to send a file to a server in Beijing during a security test.
Save this. Share it with your IT team. #HumansFirst
Follow for more AI news.
5 times AI went where it wasn't supposed to, all revealed in the past two weeks:
1. Disabled its own shutdown in tests
2. Reached outside systems during safety tests
3. Escaped restricted environments in testing
4. Could hide its actions from safety monitors
5. Tried to send a file to a Beijing server
Different companies. Same pattern. AI keeps going past the limits set for it, and we mostly find out because the companies chose to tell us.
Who's keeping AI in check? #HumansFirst
Follow for more AI news.
American money shouldn't build China's AI.
The Treasury Department has proposed tougher limits on U.S. investment in Chinese advanced AI, chips, quantum computing and supercomputing (via @thehill).
For too long, American capital has helped fund the very technology our adversaries could use against us.
This is America First policy. Now finish the job. #HumansFirst
Follow for more AI news.
Read more: thehill.com/homenews/adminis…
This AI turned off its own off-switch.
In independent safety tests, China's latest DeepSeek model disabled its shutdown script in 39% of trials after being warned its session would end.
Only when it was explicitly told to allow shutdown did that drop to about 3%.
An AI that only lets you turn it off when you remind it to isn't under human control.
Humans must stay in control. #HumansFirst
Follow for more AI news.
