You've probably heard us say this countless times: GPT-3, the gargantuan AI that spews uncannily human-like language, is a marvel. You can tell with a simple trick: Ask it the color of sheep, and it will suggest "black" as often as "white"--reflecting the phrase "black sheep" in our vernacular. That's the problem with language models: because they're only trained on text, they lack common sense. Now researchers from the University of North Carolina, Chapel Hill, have designed a new technique to change that. They call it "vokenization," and it gives language models like GPT-3 the ability to "see."

You've probably heard us say this countless times: GPT-3, the gargantuan AI that spews uncannily human-like language, is a marvel. You can tell with a simple trick: Ask it the color of sheep, and it will suggest "black" as often as "white"--reflecting the phrase "black sheep" in our vernacular. That's the problem with language models: because they're only trained on text, they lack common sense. Now researchers from the University of North Carolina, Chapel Hill, have designed a new technique to change that. They call it "vokenization," and it gives language models like GPT-3 the ability to "see."

AI technology has become increasingly sophisticated in recent years. So many products and services now rely on the technology to provide automation and intelligence that it is deeply and irrevocably intertwined with our everyday world. Whether through devices we use to enable convenience at home or in the way products we use all the time are manufactured, its impact is everywhere, driving innovation in just about every aspect of our lives. But there are missing pieces to this puzzle that still cause frustration for end-users and present significant challenges for researchers trying to improve how AI technology performs. A common sense approach Before his passing in 2018, Microsoft co-founder Paul Allen dedicated an admirable amount of time and resources to solving an essential challenge that seems to come up again and again: The fundamental lack of common sense in AI technologies.

Paul Allen is pouring more money into artificial intelligence research. The Microsoft co-founder on Wednesday announced Project Alexandria, a new initiative that aims to help develop "common sense AI." Allen is committing $125 million over the next three years to the project, housed at the Seattle-based Allen Institute for Artificial Intelligence (AI2). The money will go toward Project Alexandria, as well as other work being done at AI2. AI research has seen exponential advancements over the past decade, but machines are still incapable of certain types of problem solving or dealing with unpredictable situations. "When I founded AI2, I wanted to expand the capabilities of artificial intelligence through high-impact research," Allen said in a statement. "Early in AI research, there was a great deal of focus on common sense, but that work stalled. AI still lacks what most 10-year-olds possess: ordinary common sense. We want to jump start that research to achieve major breakthroughs in the field."