A brand new nationally representative survey has found Australians are deeply concerned concerning the risks posed by artificial intelligence (AI). They want the federal government to take stronger measures to make sure secure development and use.

We conducted the survey in early 2024 and located that 80% of Australians consider stopping catastrophic risks through advanced AI systems must be a worldwide priority comparable to pandemics and nuclear war.

As AI systems develop into more powerful, decisions about how we develop, deploy and use AI are critical. The promise of powerful technology could entice firms – and countries – to accomplish that Race ahead without listening to the risks.

Our findings also reveal a spot between the AI ​​risks that media and government are likely to deal with and the risks that Australians consider most vital.

Public concern about AI risks is growing

The development and use of increasingly powerful AI continues to advance. Current publications resembling Google’s twins And Anthropics Claude 3 have seemingly near-human abilities in skilled, medical and legal areas.

However, the hype has been tempered by increasing public and expert concern. Last yr, greater than 500 people and organizations submitted applications to the Australian Government Discussion paper on secure and responsible AI.

They described AI-related risks resembling biased decision-making, lack of trust in democratic institutions resulting from misinformation, and increasing inequality resulting from AI-related unemployment.

Some even fear that a very powerful AI may lead to this a worldwide catastrophe or Extinction of humanity. Although this concept is hotly debated in quite a few cases three large SurveyMost AI researchers consider that there’s no less than a 5% likelihood of superhuman AI being “extremely bad (e.g. causing human extinction).”

The potential advantages of AI are significant. AI is already resulting in this Breakthroughs in biology and medicineand it’s used to Control of fusion reactors, which could in the future provide carbon-free energy. Generative AI improves productivityespecially for learners and students.

However, the speed of progress raises alarm bells. People fear that we’re unprepared to cope with powerful AI systems that may very well be misused or behave in unintended and harmful ways.

In response to such concerns, world governments are attempting to control the situation. The European Union has agreed to a draft AI law, the United Kingdom has founded an AI security institutewhile US President Joe Biden recently signed an executive order to put it on the market safer development and control of advanced AI.

Australians want measures to forestall dangerous consequences of AI

To understand how Australians take into consideration AI risks and the best way to address them, we surveyed a nationally representative sample of 1,141 Australians in January and February 2024.

We have found that stopping “dangerous and catastrophic consequences of AI” is the highest priority for presidency motion for Australians.

Australians are most concerned about AI systems which are unsafe, untrustworthy and inconsistent with human values.

Other major concerns include using AI in cyberattacks and autonomous weapons, AI-related unemployment, and AI failures causing damage to critical infrastructure.

Strong public support for a brand new AI regulator

Australians expect the federal government to take decisive motion on their behalf. An overwhelming majority (86%) desire a latest government agency dedicated to AI regulation and governance, just like this Management of therapeutic goods for medication.

Nine in ten Australians also consider the country should play a number one role in international efforts to control AI development.

Perhaps most strikingly, two-thirds of Australians would support a six-month pause in AI development to permit regulators to catch up.

Government plans should reflect public expectations

In January 2024, the Australian government published a Interim plan to handle AI risks. This includes strengthening existing laws on data protection, online security and disinformation. It can be recognized that our current regulatory frameworks aren’t sufficient.

The preliminary plan calls for the event of voluntary AI safety standards, voluntary labeling of AI materials and the establishment of an advisory board.

Our survey shows Australians support a more safety-focused approach that puts regulation first. This is in contrast to the targeted and voluntary approach of the interim plan.

It is a challenge to encourage innovation while stopping accidents or misuse. But Australians would favor the federal government to prioritize stopping dangerous and catastrophic consequences over “delivering the advantages of AI to everyone”.

Some ways to do that are::

  • Establish an AI security laboratory with the technical capability to check and/or monitor probably the most advanced AI systems

  • Establishment of a dedicated AI regulatory authority

  • Defining robust standards and guidelines for responsible AI development

  • requires independent auditing of high-risk AI systems

  • Ensuring corporate liability and redress for AI damages

  • Increasing public investment in AI security research

  • Actively engage the general public in shaping the longer term of AI governance.

Figuring out the best way to effectively control AI is considered one of them The great challenges facing humanity. Australians are keenly aware of the risks of failure and need our government to handle this challenge immediately.

This article was originally published at theconversation.com