In light of the tremendous leaps of artificial intelligence and its extended uses between positive and negative, global safety indicators have emerged in artificial intelligence, including: the principles of the Organization for Economic Cooperation and Economic Development «OECD), the AI Ethics Guidelines initiative, the European Union indicators and the General Data Protection Regulations« GDPR), and the document of the framework for adopting artificial intelligence as a guide.
The Future of Life Institute «FLI) has published the artificial intelligence integrity index, which was built on the evaluation of safety efforts in 7 advanced leading companies with the aim of stimulating a“ race towards the summit ”in safety and transparency standards, where companies were evaluated in 6 basic areas that included 32 indicators, are: Risk evaluation“ such as dangerous capabilities tests ”, current damage“ protection from penetrations and privacy ”, safety frameworks, safety frameworks, safety frameworks, safety frameworks, safety frameworks, safety frameworks, safety frameworks, safety frameworks, safety frameworks, safety frameworks “Technical controls and governance”, existential safety “planning AGI/ASI”, accountability and governance “institutional structure and reporting policies”, and participation of information “cooperation with governments and researchers”.
Companies were classified from A “higher” to F “lowest”, and the main results showed great safety gaps, so that there is no company ready for the risks of public intelligence “AGI”. All companies got D or less in the field of existential safety, although their leaders expect a general intelligence to build a decade as the committee described this contradiction as “worrying”. There are only 3 “Anthropic, ONAI, Google” companies that conducted tests for the dangers of biology or cyber terrorism, but their methodologies lacked transparency and external testing. Oben AI is only the one who has published a full policy to report violations. As for the rest, either the lack of policy or imposing restrictions on employees.
It also turned out that the gap expides the technical development and safety controls through the race towards launching stronger models “such as Grok-4 from X.AI”, which exceeded safety improvements in the absence of mandatory regulatory standards, while Chinese companies have obtained an estimate F, the report did not address cultural and legal differences, for example; China has governmental systems to regulate artificial intelligence, while Western standards “such as transparency in governance” are less prominent in the Chinese framework!
The index made recommendations to all companies, including publishing a preliminary plan for the mechanism of controlling AGI systems and its allocated recommendations, such as Meta’s recommendation to enhance investment in technical safety research, and Deepseek’s recommendation to address the “Jailbreak” and Openai to rebuild the safety team after a major resignation.
The trend towards huge investments and a racing to develop super intelligence systems may lose control over us, so the world needs to reconsider an approach to the safety of artificial intelligence and capitalist motives for money harvesting, and perhaps hidden forces, digital sovereignty and influence, all of which are motivated by this unique global race.
