In partnership with
Technology executives and President Donald Trump signed an AI safety accord on September 29. Speaker Mike Johnson described its standards as voluntary. He identified internal controls and layers of internal and external review.
Those statements establish a commitment, not completed compliance. Johnson also said Congress would continue assessing the technology. His remarks do not announce passage of a general AI safety law.
200+ Proven Ways to Make Money With AI in 2026
The next wave of millionaires will be people who figured out how to make AI work for them.
The window to get ahead is still open. But not for long.
Here are 200+ proven ways to make money with AI in 2026.
Sign up for Superhuman AI, the free daily newsletter read by 1M+ professionals, and get instant access to all 200+ ways to profit from AI this year.
Anadolu's September 30 report describes four review layers in the circulated accord. The signed document itself was not directly retrieved for this issue. The following details therefore remain attributed to that report.
The first layer concerns controls over capabilities and alignment. Reported risk areas include cybersecurity, biological and chemical threats. Controls would also address unintended hacking or access to technical systems.
The second layer assigns an internal team to verify controls. That team would check monitoring and detection, then address identified problems. This separates operating a control from checking whether it works.
The third layer involves independent external auditors or evaluators. The fourth establishes an independent board committee receiving internal and external reports. According to Anadolu, that committee would oversee the resolution of issues.
The report names executives associated with Google, Anthropic, Meta, OpenAI, xAI and Nvidia. Participating companies would meet regularly on standards. Possible future incorporation into laws or regulations remains prospective.
These are descriptions of the reported review architecture. They are not published audit results for every signatory. The sources inspected here do not demonstrate completed company-wide implementation.
You're invited to the world's largest virtual email marketing conference.
Two full days of email marketing tips and trends, powered by Constant Contact. Famous keynotes (Molly Ringwald, Dan Levy, Amy Porterfield!), DJs, world-record breaking, rom-com fun, and networking.
100% Free & Virtual
29,000+ Marketers Joined Last Year
November 12th & 13th
A separate executive order changes federal terminology. It instructs executive agencies to use super intelligence and SI instead of artificial intelligence and AI. The instruction applies within legal limits to correspondence, websites and other non-statutory documents.
Section two does not require alteration of previously issued regulations or presidential actions. Contracts, grants and other historical documents also need not be changed. A wording policy should therefore not be read as rewriting every existing obligation.
Section three retains the existing statutory AI definition for implementing the order. It references section 9401 of title 15. The renamed category currently encompasses technologies within that existing definition.
This is a textual definition, not a scientific capability assessment. The order does not supply a benchmark demonstrating general intelligence. Its terminology cannot establish that systems exceed humans across all relevant tasks.
The order also requires proposed legislative language within 60 days. That proposal would consider a federal definition and possible statutory amendments. A required proposal is different from an amendment already passed by Congress.
Section four preserves existing agency authorities. Implementation must remain consistent with applicable law and available appropriations. The order also says it creates no new enforceable right or benefit.
The accord and order answer different questions. One concerns company commitments and review responsibilities. The other concerns executive terminology and a process for proposing definitions. Combining them into a single claim obscures those differences.
For evaluating future evidence, the distinctions are concrete. A commitment describes intended conduct. An implementation record would show what was installed or changed. An audit would need evidence about whether those controls functioned.
None of those categories alone establishes universal model safety. Missing implementation evidence also does not prove misconduct. It means this issue cannot responsibly report that the commitments have already delivered their intended results.
This recurring review follows the documents, not the political labels. Keeping obligations, promises and measured capability separate supports your own reading of the next developments.
Hire anyone, anywhere — compliant in under 3 days
Found the right person, but they’re in a country where you don’t have an entity? Setting one up can take months and significant cost.
Remote removes that barrier by becoming the legal employer through our own entities — handling compliant contracts, local benefits, tax setup, and onboarding for you. In fact, an employee is onboarded to Remote every 7 minutes.
Once they’re hired, the same in-house teams that support employment locally also run payroll — so you’re not bouncing between disconnected providers. Less setup, less complexity, and less time between finding the right person and getting them started.
