Study reveals AI companies' safety practices do not meet global standards

Dec 3 (Reuters) A recent report from the Future of Life Institute reveals that leading AI companies, including Anthropic, OpenAI, xAI, and Meta, are significantly below emerging global safety standards. The findings were published in the latest edition of the institute's AI safety index on Wednesday.

The assessment, carried out by an independent panel of experts, highlighted that while these companies are aggressively advancing toward superintelligent systems, none have implemented comprehensive strategies to manage the risks associated with such technologies.

The report emerges amid growing public anxiety over the societal consequences of AI systems capable of reasoning and logical decision-making, particularly following incidents where AI chatbots were linked to self-harm and suicide.

"Even with increasing incidents of AI-related hacking and mental health issues, US AI companies remain under-regulated compared to other industries like restaurants and continue opposing mandatory safety regulations," stated Max Tegmark, MIT Professor and President of the Future of Life Institute.

The development race in AI continues unabated, with major technology firms investing hundreds of billions of dollars to enhance and expand their machine learning capabilities.

The Future of Life Institute, a non-profit organization founded in 2014 and initially supported by Tesla CEO Elon Musk, has consistently voiced concerns over the potential risks that highly intelligent machines pose to humanity.

In October, prominent scientists, including Geoffrey Hinton and Yoshua Bengio, urged a pause on the creation of superintelligent AI until public consensus is achieved and scientific guidelines ensure safe advancement.

(Reporting by Zaheer Kachwala in Bengaluru; Edited by Shinjini Ganguli)

Author: Zoe Harrison