Artificial Intelligence (AI) experts are amazed by the rapid speed of AI development, according to a recent survey. The study reveals that even the professionals perceive a significant acceleration in the field, pointing towards an imminent AI-driven automated world. The survey also uncovers the divide between 'accelerationists' and 'doomers', and examines their contrasting viewpoints on the pace of AI evolution.
AI experts attest to rapid development
A recent survey by San Francisco-based research firm AI Impacts has shed light on the perceptions of AI researchers in regards to the rapid growth of AI technologies. The survey, which was funded by billionaire and Facebook co-founder Dustin Moskovitz's grant-making entity Open Philanthropy, was based on responses from 2,778 AI researchers. The results showed that there is a sense of acceleration in the field, making it clear that the pace of AI development is not just a figment of our imagination.
Survey indicates acceleration in AI
The survey discovered a significant shift in the average prediction timeframe for various AI-related tasks. The researchers noted that the 50th percentile year for these tasks to become feasible shifted forward by a whole year between 2022 and 2023. This finding is particularly powerful as it suggests that the experts are not just acknowledging an acceleration in AI development, but are continually revising their time estimates based on observed advancements.
Sharp prediction shifts for HLMI and FAOL
Among the standout findings of the survey were the drastic shifts in aggregate forecasts for two key concepts: High-Level Machine Intelligence (HLMI) and Full Automation of Labor (FAOL). The estimated arrival time for HLMI reduced by 13 years between 2022 and 2023, while the forecast for FAOL decreased by 48 years over the same period. Such drastic shifts reflect the accelerating pace of AI development and its potential to disrupt various occupations sooner than previously expected.
The survey also explored the attitudes of researchers towards the pace of AI development. While a small group of 'doomers' advocated for a slower pace of development, 'accelerationists' were in favor of faster progress. The majority of the respondents, however, seemed to favor the status quo, with the 'somewhat slower', 'current speed', and 'somewhat faster' responses accounting for nearly 80 percent of all answers. This suggests a degree of caution and divide among the experts regarding the rapid progression of AI technologies.
It's important to note that while the survey respondents are experts in AI, they may not necessarily have expertise in forecasting the AI trajectory. As such, their predictions might not always account for non-technical factors that can influence the path of AI development. Nonetheless, their insights are valuable as they provide a peek into the collective aspirations, fears, and expectations of those at the forefront of AI innovation.