Forecasting technological innovation and its progress have been much more difficult as there is no past data to draw upon future technologies that are at best poorly understood.
AI systems capabilities have been even more difficult as the forecasting has progressed broadly though one does not yet know the fundamental architectures that drive such systems.
However, efforts to these ends are not unified and the study of AI forecasting does not appear to be directed broadly at a well-understood objective.
A framework for forecasting or modeling AI progress has come up with a new alternative that outlines the utilities in both scenario analysis techniques as well as the statistical, judgmental, and data-driven forecasting techniques.
Over the past two decades, the use of scenario analysis techniques has increased enormously for mapping complex environments, complex systems, and complex technologies. In Particular, forecasting drives to focus on three such techniques.
Firstly it is a relatively little practical obscured application method, yet has more significant potential for mapping the path of the possible future for which there are high levels of uncertainty.
The second originated to represent social scientific knowledge through directed graphs and has become a common method for scenario analysis in multi-organizational contexts.
The third extends the second by making computable methods for quantitative forecasting, and also has practical uses across various other domains in a large number of applications.
Each of these techniques offers insights that contribute to the holistic framework for forecasting transformative AI.
The necessity of planning for Human Level Artificial Intelligence (HLAI) is obvious. It is plausible and likely that AI will have severe transformative effects without reaching human-level intelligence mostly on society.
Therefore, the purpose of ensuring that AI is developed is to do the best possible for humanity which can identify the primary task of AI forecasting to forecasting transformative AI.
Transformative AI is defined as a set of AI technologies that has the potential to transform the quality of life for social groups or society in ways that dramatically reshape social structures.
Human-level artificial intelligence(HLAI) has the potential to anticipate difficult transformable society in ways. Not only its impacts are difficult to imagine, but the notion of HLAI itself is ill-defined.
Ill-defined defines what may be indicative of human-level intelligence to some may not be sufficient to others, and there is also no definitive test for human-level intelligence.
This has led studies concerned with HLAI or forecasting AI progress to focus on the replacement of humans at tasks or jobs. The lack of an objective for HLAI is due in part to the fact of how to create it.
Theoretically, HLAI could be instantiated by one algorithm or developed and constructed by combining different components.
The other unique challenges faced in forecasting HLAI have adequately addressed the methods that integrate a variety of possible paths or diverse information paths that are necessarily required.
By assessing progress, particularly, the rate of progress is essentially needed when developing any type of technology forecasting model.
This is because the naïve assumption and its historical trends are extrapolated to the future correctly many times, and, consequently, trend extrapolation has been a very powerful forecasting technique.
Variable Indicators can be used for extrapolation or for building statistical forecasting models as they are good for predicting future progress.