A recent report, AI 2027, has presented an ambitious timeline regarding artificial intelligence advancements. According to this projection, artificial general intelligence may become achievable by 2027, promptly succeeded by superintelligence in 2028. The roadmap details an intelligence explosion, where AI systems begin independently accelerating their research and development capabilities.
Notable predicted milestones include the emergence of an AI coder surpassing human capabilities at the beginning of 2027, followed by the public deployment of artificial general intelligence (AGI) around midyear, and finally, the advent of an AI researcher exceeding human ability later within that same year.
The report considers two distinct possible outcomes: one characterized by a perilous, competitive push toward these milestones with a heightened risk of catastrophic results, and another scenario depicting a more controlled and deliberate approach aimed at ensuring a safer and more regulated transition. This analysis highlights potential dangers, emphasizing how misalignment risks may sharply increase within the context of an intense international race toward AI supremacy.