Feds move to regulate automated decision-making - iTnews

The federal government has decided that automated decision-making needs a good, firm regulating. And who better to lead this charge than the Attorney-General? Because nothing says 'nimble, tech-savvy oversight' like a lawyer tasked with writing rules for algorithms that probably already have a head start. This comes hot on the heels of the 'world-first' AI standards from the same government, which we noted last time were all fanfare and no substance. Now they're proposing a framework to regulate ADM within federal agencies — which is a bit like installing a speed bump after the car has already crashed into the tree.

The Attorney-General will develop a framework to regulate ADM. Great. Because the same government that can barely manage its own IT systems (remember the Robodebt debacle? That was automated decision-making at its finest) is now going to write the rulebook. The 'safety priorities' are mentioned, but let's be honest: safety priorities in government AI are about as reliable as a chocolate teapot. The only priority here is to be seen doing something before the next scandal erupts.

And the scope? Just federal agencies. Not the private sector, not the social media platforms, not the credit-scoring algorithms that already run most of our lives. Just the nice, controllable, public-sector sandbox. Because we all know the real problems with AI are in government tax forms, not in insurance, hiring, or policing. Ahem. So yes, regulate away. Write the framework. Hold the consultations. And in three years, when the first real crisis hits, we can all admire the beautiful, unused architecture.