In a bold move that signals the accelerating pace of artificial intelligence development, OpenAI and Microsoft are reportedly preparing to test artificial general intelligence (AGI) systems in real-world scenarios as early as 2025. According to Business Insider, both tech giants are developing frameworks to evaluate whether their AI systems can outperform humans across various practical tasks. This aggressive timeline suggests that AGI—AI systems with human-level intelligence across multiple domains—may be closer to reality than many experts previously predicted.
The testing plans reveal a significant shift in how these companies view their AI progress. Rather than focusing solely on benchmark tests or controlled environments, OpenAI and Microsoft aim to deploy their systems in genuine real-world applications, potentially including software development, scientific research, and business operations. This approach would provide more meaningful measurements of AI capabilities while simultaneously demonstrating practical value. The competition between these tech powerhouses appears to be intensifying, with each company racing to achieve the historic milestone of creating AI that can match or exceed human performance across diverse cognitive tasks.
Source: https://www.businessinsider.com/openai-microsoft-duel-agi-real-world-tests-ai-better-humans-2025-6