Possibly Harmful Smart Software: Heck, Release It

July 13, 2020

So, what changed? A brief write-up at Daijiworld reports that the “Elon Musk-Founded OpenAI Releases Text Tool it Once Called Dangerous.” This software can rapidly generate fake news that is so believable its makers once deemed it too dangerous to be released. Now, though, OpenAI seems to have had a change of heart. The API is being released in a private beta rather than into to the world at large as a test run of sorts. Citing an OpenAI blog post, the write-up reveals:

“‘In releasing the API, we are working closely with our partners to see what challenges arise when AI systems are used in the real world,’ OpenAI said in a blog post last week. ‘This will help guide our efforts to understand how deploying future AI systems will go, and what we need to do to make sure they are safe and beneficial for everyone.’ The API that OpenAI finally decided to release provides a general-purpose ‘text in, text out’ interface, allowing users to try it on virtually any English language task. Interested buyers can integrate the API into their product and develop an entirely new application. ‘Given any text prompt, the API will return a text completion, attempting to match the pattern you gave it. You can “program” it by showing it just a few examples of what you’d like it to do; its success generally varies depending on how complex the task is,’ OpenAI said.”

Where will it go from here—will OpenAI decide a general release is worth the risk? We’re guessing it will. Evidently this software is just too juicy to keep under wraps.

Cynthia Murrell, June 25, 2020

Comments

Got something to say?





  • Archives

  • Recent Posts

  • Meta