Submission + - New AI text generator may be too dangerous to release
umafuckit writes: The Guardian reports that GPT-2, the new text generator from OpenAI, is so good that it risks being used to malicious ends. As a result, OpenAI are delaying release. The model underlying GPT2 has been trained on 40 GB of text data and is able to write highly realistic prose. For example, after being fed the first few paragraphs of a news story, the AI continues with the paragraph: "Asked to clarify the reports, a spokesman for May said: ‘The PM has made it absolutely clear her intention is to leave the EU as quickly as is possible and that will be under her negotiating mandate as confirmed in the Queen’s speech last week.’”
Is this text generator truly groundbreaking or is the delay in release just a tactic to attract attention to it?
Is this text generator truly groundbreaking or is the delay in release just a tactic to attract attention to it?
