Hello World : Meet Generative AI | Amazon Web Services
WATCH NOW
Great care needed to deal with current limitations With human-level performance on various professional and academic benchmarks , the latest version of OpenAI ’ s GPT technology , GPT-4 , is highly impressive .
Despite its capabilities , however , GPT-4 has similar limitations as earlier GPT models . “ Most importantly , it still is not fully reliable ,” the OpenAI team says . “ Great care should be taken when using language model outputs , particularly in high-stakes contexts , with the exact protocol ( such as human review , grounding with additional context , or avoiding high-stakes uses altogether ) matching the needs of a specific use-case .”
Aaron Kalb , Chief Strategy Officer and Co-Founder at Alation , reiterates this final point : tools like GPT should not yet be trusted to advise on important decisions .
“ That ’ s because it ’ s designed to generate content that simply looks correct with great flexibility and fluency , which creates a false sense of credibility and can result in so-called AI ‘ hallucinations ’.
As Kalb – who , when working at Apple , was part of the founding team behind its groundbreaking Siri voice assistant – explains , the authenticity and ease of use that makes GPT so alluring is also its most glaring limitation : “ Only if and when a GPT model is fed knowledge with metadata context – so essentially contextual data about where it ’ s located , how trustworthy it is , and whether it is of high quality – can these hallucinations or inaccurate responses be fixed and GPT trusted as an AI advisor .”
“ GPT is incredibly impressive in its ability to sound smart . The problem is that it still has no idea what it ’ s saying . It doesn ’ t have the knowledge it tries to put into words . It ’ s just really good at knowing which words ‘ feel right ’ to come after the words before , since it has effectively read and memorised the whole internet . It often gets the right answer since , for many questions , humanity has collectively posted the answer repeatedly online .” aimagazine . com 45