- A man uses artificial intelligence (AI) to create a chatbot that mimics his late fiancé.
- The groundbreaking artificial intelligence technology was designed by Elon Musk’s research group OpenAI.
- OpenAI has long warned that the technology can be used for large-scale information campaigns.
- For more stories, please visit www.BusinessInsider.co.za.
After Joshua Barbeau’s fiancé passed away, he talked to her for a few months. Or, to be more precise, he was talking to a chatbot programmed to sound exactly like her.
In the story of the San Francisco Chronicle, Barbeau introduced Project December in detail, which is a software that uses artificial intelligence technology to create a surreal chatbot, how it reproduces the experience of talking with the deceased fiance. All he has to do is to insert old messages and provide some background information, and then the model can suddenly imitate his partner with amazing accuracy.
This may sound like a miracle (Or Black Mirror episode), But the creators of artificial intelligence warn that the same technology can be used to promote large-scale misinformation campaigns.
Project December is supported by GPT-3, an artificial intelligence model designed by OpenAI, a research group supported by Elon Musk. By using a large number of artificial text data sets (Reddit threads are especially helpful), GPT-3 can imitate human writing and produce everything from academic papers to letters from former lovers.
This is by far the most complex and dangerous language-based artificial intelligence programming.
When OpenAI released GPT-2 (the predecessor of GPT-3), The group wrote It may be used in a “malicious way”. The organization expects that bad actors using the technology can automate “abusive or fake content on social media,” “generate misleading news articles,” or “impersonate others online.”
The organization wrote that GPT-2 can be used to “unlock new capabilities for these participants that have not been anticipated.”
OpenAI staggered the release of GPT-2 and still restricted access to the superior GPT-3 in order to “give people time” to understand the “social impact” of such technologies.
Misinformation is already rampant on social media, even though GPT-3 is not widely used. A new study finds YouTube’s algorithm Still pushing for misinformation and non-profit anti-digital hat centerse Recently identified 12 responsible persons 65% of Covid-19 conspiracy theories are shared on social media. They are known as the “twelve people with false information” and have millions of fans.
With the continuous development of artificial intelligence, Oren Etzioni, CEO of the Allen Institute’s non-profit biological science research group, said, Tell the insider before It will only become harder to tell what is real.
“The question is,’Is this text, picture, video or email true?’ It will become increasingly difficult to answer based on the content alone,” he said.
