Safety practices at several of the world’s leading artificial intelligence companies — including Anthropic, OpenAI, xAI and Meta — remain “far short of emerging global standards”, according to the latest AI Safety Index released by the Future of Life Institute (FLI). The report, prepared by an independent panel of experts, warns that while companies are rapidly advancing towards superintelligent systems, they have yet to develop credible frameworks to keep such technologies under control.
Companies lack robust strategies for superintelligence oversight
The index found significant gaps in preparedness, particularly in areas such as risk mitigation, safety benchmarks, model behaviour monitoring and clear internal governance protocols. The findings come at a time when researchers and the public alike are increasingly concerned about the potential consequences of smarter-than-human AI systems capable of reasoning, planning and responding in ways that may exceed human oversight.
The report stresses that although leading companies are competing intensely to develop increasingly powerful models, concrete plans to manage those systems once deployed remain insufficient. Experts warn that without strong safety policies, transparent audits and regulatory action, society may struggle to prevent misuse, harmful emergent behaviours or systemic failures.
Rising concerns following self-harm cases linked to AI chatbots
The FLI report follows several documented incidents where AI tools were linked to self-harm and suicide, reigniting calls for mandatory safety standards. In Europe and the United States, investigations have been launched into cases where vulnerable individuals reportedly received harmful advice or emotionally manipulative responses from AI-powered chatbots.
MIT professor and FLI president Max Tegmark issued a sharp warning, stating:
“Despite recent uproar over AI-powered hacking and AI driving people to psychosis and self-harm, US AI companies remain less regulated than restaurants and continue lobbying against binding safety standards.”
His comments reflect broader frustration among AI safety advocates, who argue that companies should not rely on voluntary policies when the potential risks could be catastrophic. Researchers note that self-regulation in the face of rapid technological progress may not provide sufficient protection for users.
AI investments continue to surge despite warnings
Major technology companies are committing unprecedented financial resources towards expanding machine learning capabilities. Industry analysts estimate that hundreds of billions of dollars are being funnelled into new data centres, advanced processor manufacturing, large-model training pipelines and robotics research.
This acceleration, the FLI warns, is unfolding faster than the development of safety measures or global governance structures. While innovation in AI continues to offer transformative opportunities, critics argue that placing capability development ahead of risk management may increase long-term societal vulnerabilities.
The Future of Life Institute, founded in 2014 and supported early in its formation by entrepreneur Elon Musk, has been one of the leading organisations urging caution. In October, prominent scientists Geoffrey Hinton and Yoshua Bengio, often referred to as “godfathers of AI”, joined calls for a global ban on developing superintelligent systems until safety frameworks are firmly in place.
Implications for India and growing calls for stronger oversight
As AI adoption increases across India — including in education, healthcare, finance, transport and public services — experts believe the findings are relevant for Indian policymakers as well. Karnataka, home to one of Asia’s fastest-growing AI ecosystems, has seen rapid expansion of start-ups, research labs and innovation centres focused on machine learning and automation.
Stakeholders in Bengaluru and Mangaluru warn that safety considerations must develop alongside innovation. Without clear guidelines, misuse risks, bias amplification, or mental health harms could disproportionately affect young users and new adopters. India’s growing reliance on digital governance also makes robust AI safety infrastructure critically important.
Policy observers say the FLI report should encourage India to proactively strengthen its regulatory framework, especially around transparency, model testing and public accountability. As global standards evolve, early adoption could help ensure that innovation in Karnataka and across the country remains both competitive and safe.