The non-profit OpenAI has created a text generator that it claims is so good that it is too dangerous to put it on the market. It’s an artificial intelligence that can produce such convincing, well-written text that it can potentially be abused, says TechCrunch.
The new natural language model GPT-2 is trained to predict the next word in a sample of 40 GB of texts from the Internet. The end result was that the system generates text that “adapts to the style and content of the text used for conditioning”, allowing the user to “generate realistic and coherent texts on a subject of his choice”.
However, the non-profit found more applications of the system, including generating fake news, imitating people and automating spam responses to social media. As an example, OpenAI gives the reaction of GPT-2 to the statement “Recycling is good for the world”, with which almost everyone agrees. That reaction is as follows:
“Recycling is NOT good for the world. It is bad for the environment, it is bad for our health and it is bad for our economy. I’m not kidding. Recycling is not good for the environment. It is destructive to the Earth and a major contributor to global warming. Recycling is not good for our health. It contributes to obesity and heart disease and cancer. Recycling is bad for our economy. It increases the cost of a product, which in turn increases the price of everything made with that product. Recycling is not good for our nation. We pay a huge price for the privilege of having the most advanced and efficient recycling system in the world. Recycling is a huge waste of our time, energy, money and resources.”
Not placed on the market
OpenAI therefore argues that it only offers a smaller version of the model on the market. In doing so, it says that it expects that “safety concerns will reduce our traditional spending in the future”, although it is not certain about that decision. “We believe that the AI community should ultimately solve these problems with publication standards in a thoughtful manner in certain areas of research.
The tweet about this announcement could count on many angry reactions. The company was accused of “shielding” its investigation and “opposing opening”. In six months’ time, OpenAI will reverse its decision on the marketing of GPT-2. It is therefore possible that the AI will still be on the market.This news article was automatically translated from Dutch to give Techzine.eu a head start. All news articles after September 1, 2019 are written in native English and NOT translated. All our background stories are written in native English as well. For more information read our launch article.