President Donald Trump and leaders from some of America's largest artificial intelligence companies have signed a voluntary agreement establishing new safety controls for advanced AI systems. Google, Anthropic, Meta, Nvidia, OpenAI and Elon Musk's xAI agreed to implement four layers of controls and audits — but the deal is not legally binding, and critics are already asking who holds the industry accountable if it polices itself.
Watch the full news report below, then read the breakdown.
Watch on YouTube: https://www.youtube.com/watch?v=KmqIHnUZniM
The four layers of safety controls
Under the agreement, the AI giants committed to:
1. Internal monitoring — tracking the behaviour and risks of their own advanced systems.
2. Dedicated internal oversight — dedicated teams responsible for AI safety inside each company.
3. Independent external evaluations — outside experts testing and auditing the systems.
4. Independent board-level oversight — company boards formally accountable for AI safety.
Legally binding? No. "Morally binding," says Trump
The agreement carries no legal force. President Trump described it as "morally binding" and said there is a strong belief that AI companies should regulate themselves rather than be regulated by government. The document acknowledges that its safety measures could eventually be incorporated into laws or regulations — which leaves the door open, but for now the industry is on the honour system.
Even signatories have doubts
Anthropic CEO Dario Amodei, who has previously called for government intervention to address AI risks, signed the agreement while saying questions remain about how those risks should ultimately be addressed. It is a telling sign: the people building the most advanced AI systems are not certain self-regulation is enough.
The other battle: who pays for AI's power hunger
The debate now extends beyond the technology itself to the enormous data centres needed to run advanced AI. US lawmakers have been considering legislation to prevent ordinary electricity customers from bearing the infrastructure costs created by massive new data centres.
The proposed Ratepayer Protection Act would require state utility regulators to make data centres and other large electricity users responsible for the incremental costs of the generation, transmission and distribution upgrades needed to serve them. The measure passed the House of Representatives overwhelmingly, 417-3, but failed to advance in the Senate on a 57-43 vote — three votes short of the 60 needed.
What it means for the rest of the world
For workers, creators and businesses everywhere, two things stand out. First, AI is now firmly a policy issue at the highest level — the pace of regulation will shape how fast AI reshapes jobs and industries. Second, self-regulation means the safety of the AI tools you use every day currently depends largely on the companies themselves. The smart move remains the same: use AI tools actively, but stay informed and keep building skills no machine can replace.
Source: KFDM / Fox4 Beaumont — YouTube channel: https://www.youtube.com/@newsonkfdm
