On the 5th, the United States, the European Union, and the United Kingdom signed the world’s first multinational agreement on the use of artificial intelligence (AI), claiming it to be legally binding. At the same time, Australia also plans to impose "guardrail" measures on high-risk AI development, including requiring human intervention and transparency, showing that multiple countries are taking concrete actions to address the potential risks of AI.
The three major Western jurisdictions developing AI technology— the US, EU, and UK—signed the Council of Europe's AI Convention on the 5th, emphasizing human rights and democratic values as they seek to regulate both public and private sector systems.
The convention, drafted over two years by more than 50 countries, including Canada, Israel, Japan, and Australia, is legally enforceable, requiring signatories to take responsibility for any harmful or discriminatory outcomes from AI systems. It also mandates that the outputs of such systems must respect fairness and privacy, and that victims of AI-related infringements have legal means to seek redress.
Peter Kyle, the UK Minister for Science, Innovation, and Technology, stated that for such a rapidly advancing innovation like AI, taking the first step towards a multinational agreement is important. "This is the first global agreement with substantial binding force, and it brings together very different countries."
Hanne Juncher, who led the negotiations for the Council of Europe, noted that at the time of signing on the 5th, ten participating countries are expected to be the first to ratify the convention. "This confirms that the convention goes far beyond Europe. These signatory countries have been highly engaged in negotiations and are satisfied with the outcome."
Although this agreement claims to be legally binding, critics argue that it lacks sanctions such as fines and primarily relies on relatively weak enforcement through monitoring to evaluate businesses.
On the 5th, Australia’s Minister for Industry and Science, Ed Husic, also announced ten new voluntary guidelines for AI systems, effective immediately, with a one-month public consultation period to determine whether high-risk AI development should be required to adhere to them in the future.
Additionally, according to a report by AFP cited by the Central News Agency, the International Labour Organization (ILO) has stated that workers’ wealth is shrinking globally by trillions of dollars each year. In its periodic report on global employment, the ILO analyzed data from 36 countries and found that from 2019 to 2022, the global labor income share declined by 0.6 percentage points, and this trend has since stabilized. Though 0.6 percentage points may seem small, it actually represents an annual income gap of nearly $2.4 trillion, with AI posing a threat to wages and potentially widening inequality.
The study highlights that the COVID-19 pandemic is the primary factor behind the decline in workers' income, with almost half of the income loss occurring between 2020 and 2022. During this period, explosive technological advancements, including automation, were major contributors to the reduction in the global labor income share.
The ILO noted that workers currently account for only 52.3% of global income, with the remainder going to capital income—held by owners of assets such as land, machinery, buildings, and patents.