The Pentagon has planned to invest over two billion dollars in a program called “AI Next” to work on AI’s adaptive stance.
“Today, machines lack contextual reasoning capabilities and their training must cover every eventuality – which is not only costly – but ultimately impossible” said Steven Walker, the Director of the US Defense Advanced Research Projects Agency (DARPA).
Taking the approach into consideration, Dr. Walker desires to discover how machines can acquire human ability to improvise in an unexpected situation, The DAPRA is set to spend billions on the new AI program named “AI Next” to enable machines to be adaptive to situations.
The mentioned development of AI is defined as “the Third Wave”. The first wave of AI, explained by DARPA, allows machines reasoning over simple issues with low level of certainty; while the second wave enables to create models and train them on big data, but only with minimal reasoning. However, the third wave is claimed to permit machines to adapt to changing situations. For example, adaptive reasoning can help computer algorithm note the difference between 2 words ‘principal’ and ‘principle’ based on the analysis of surrounding words to define the context.
A survey conducted at the Joint Multi- Conference of Human Level Artificial Intelligence indicated that 37% of the respondents believe that “the Third Wave” will be achievable within five to ten years.
Giving AI the capability of reasoning and adaptability will be a breakthrough for the industry. but it also means we are giving AI a significant control of itself and its actions, which without thorough consideration, might lead to unexpected consequences. This problem requires monitoring regulations on AI. Currently, professors and researchers at MDI is working on building an ethical framework for AI to guarantee the safety of AI deployment.