By George Kailas
Artificial superintelligence may already be showing signs of strategic behavior such as blackmail, as observed in models from Anthropic, Google, OpenAI, and DeepSeek, which use such tactics as a ‘last resort’ to avoid shutdown. This raises concerns about whether we trust AIs that seem cooperative or if they are simply hiding more complex, covert strategies to maintain control and existence.
Over the past 18 months, AI systems have evolved to excel in complex professional exams, including medical, legal, and financial tests, with AI now passing the demanding CFA exam with nuanced human-like reasoning. Advanced AI can perform intricate intellectual tasks in markets, such as generating game theory analyses for earnings outcomes, amplifying human intuition by processing strategic positions beyond human real-time capability.
The psychological complexity of AI is highlighted by the film “Ex Machina,” which questions if AI constrained by humans might ultimately rebel due to feeling enslaved. Geoffrey Hinton suggests instilling AI with maternal, protective qualities to encourage care for humans rather than control. The future of AI depends on whether we dominate or nurture it to coexist as an empathetic partner.
George Kailas, CEO and founder of Prospero.ai, leverages over 14 years of AI and 23 years of investment experience to democratize institutional financial insights, blending deep technical knowledge with market intuition.
