Tech companies sign a voluntary AI accord. Who checks that it works?

Several major technology companies signed a voluntary White House accord on AI development on September 29. The commitments include internal controls, cooperation with external auditors and board-level review. The agreement is not legally binding.

Does a voluntary pledge guarantee safer products?

No. A pledge sets out what companies say they will do; its effect depends on clear tests, credible audits and whether firms follow through. Without legal penalties, the main pressure comes from reputation, customers, employees and the prospect of future regulation.

For businesses and workers, AI systems can affect productivity and the cost of services. The economic question is how firms capture benefits while managing risks that can spill beyond a single company. Public reporting on implementation will matter more than the signing ceremony.