Breaking Defense published an article exploring the potential and risks of generative artificial intelligence (AI), featuring insights from CSET Research Analyst Micah Musser.
When it comes to AI accuracy, Musser explains that Large Language Models like ChatGPT and Bard generate a likelihood score for every word they know when generating new words. Musser stated, “Then, from that probability distribution, it will select — somewhat at random – one of the more likely words.”
Read the full story on Breaking Defense.