Tuesday 17 November, Power & Policy

File image: Aerial black-and-white view of pedestrians walking on cobblestone plaza

Study finds AI companies follow only half of safety pledges

A 2025 study found evidence that voluntary AI safety commitments were followed only 53% of the time.

Reported by The ConversationClaims checked against the source and corrected before publish, 46 days ago.

Why it matters

Empty promises won't slow AI's risks. Real enforcement and funding are needed to make safety declarations meaningful.

In Australia. Australian government websites have been targeted by AI hack attempts. Australia's AI Safety Institute was announced but receives only one-sixteenth the funding of its UK counterpart.

Key facts

53 per cent
Pledge compliance rate
Trump and six tech companies
Accord signatories
22
World leaders signing UN declaration
Three months
Notification delay
  • Trump and six tech companies signed the White House Accord on Super Intelligence this week.
  • The accord asks tech companies to regulate themselves on AI safety, a familiar promise.
  • A UN declaration from 22 world leaders called for an international AI safety regulator.
  • A decade of similar pledges has delivered little real change or rules.
  • The Partnership on AI, formed ten years ago, made the same pledges.
  • A 2025 study found that voluntary AI safety commitments were followed just 53% of the time.
  • For pledges to work, they need funding, real institutions, enforceable rules, and specific pledges instead of broad statements.
Petrol todayFill up or wait? Today's call for every capital

Today on InSnip

  1. Man dies in shooting at Belmont Local Court south of NewcastlePower & Policy, 7NEWS, 2 h ago
  2. David Jones suppliers face months of limbo on $419m debtMoney & Work, The Sydney Morning Herald, 2 h ago
  3. One Nation vows Victoria ban on puberty blockers after states reject guidancePower & Policy, SBS News, 5 h ago

Today's ten are ready.

Read today's tenAbout seven minutes, free