A survey of 2,778 AI researchers, published in top-tier AI venues, asked about the pace of AI progress and the impacts of advanced AI systems. The results suggest a 50% chance of several milestones being achieved by 2028, including autonomously building a payment processing site, creating indistinguishable songs, and fine-tuning large language models. By 2047, there is a 50% chance AI systems could outperform humans in all tasks, earlier than a 2022 survey. The chance of all human occupations becoming automatable is forecasted to reach 10% by 2037 and 50% by 2116. Most respondents expressed uncertainty about AI's long-term value, with 68.3% believing good outcomes from superhuman AI are more likely than bad. However, 48% of net optimists gave at least a 5% chance of extremely bad outcomes, and 59% of net pessimists gave 5% or more to extremely good outcomes. Over half of respondents believed substantial concern was warranted about six AI-related scenarios, including misinformation, authoritarian control, and inequality. There was disagreement on whether faster or slower AI progress would be better for humanity, but broad agreement that AI risk research should be prioritized. The survey found that AI progress is likely to accelerate, with many milestones expected within the next decade. The results highlight the uncertainty and complexity of AI's future, emphasizing the need for careful consideration of its risks and benefits.A survey of 2,778 AI researchers, published in top-tier AI venues, asked about the pace of AI progress and the impacts of advanced AI systems. The results suggest a 50% chance of several milestones being achieved by 2028, including autonomously building a payment processing site, creating indistinguishable songs, and fine-tuning large language models. By 2047, there is a 50% chance AI systems could outperform humans in all tasks, earlier than a 2022 survey. The chance of all human occupations becoming automatable is forecasted to reach 10% by 2037 and 50% by 2116. Most respondents expressed uncertainty about AI's long-term value, with 68.3% believing good outcomes from superhuman AI are more likely than bad. However, 48% of net optimists gave at least a 5% chance of extremely bad outcomes, and 59% of net pessimists gave 5% or more to extremely good outcomes. Over half of respondents believed substantial concern was warranted about six AI-related scenarios, including misinformation, authoritarian control, and inequality. There was disagreement on whether faster or slower AI progress would be better for humanity, but broad agreement that AI risk research should be prioritized. The survey found that AI progress is likely to accelerate, with many milestones expected within the next decade. The results highlight the uncertainty and complexity of AI's future, emphasizing the need for careful consideration of its risks and benefits.