Eye on AI - April 22nd, 2022
Welcome to Aigora's "Eye on AI" series, where we round up exciting news at the intersection of consumer science and artificial intelligence!
We’re delving into art and nature this week, beginning with a story about how a human used the GPT-3 language model to co-author a love novel, then switching gears to focus on how researchers are using AI to prevent volcanic disasters
GDT-3 Co-Authored Love Novel Blends the Lines of Human and AI Creation
Our lead story this week is about how a novel was co-authored with AI is yet another example of how AI is integrating into the arts, a simple story with major implications. According to the article “An Artificial Intelligence Author Makes Its Way into Literature with a Love Story,” AI has, for one of the first known times, been used to co-create a published novel. The story, written by the novelist and AI expert Dr. García-Crespo and published by AI-Creative, a think-tank agency tracking and reporting the impact of AI in the arts, was co-authored by the natural language processing AI GPT-3 and deals with a subject that’s all too human: love.
“With more than 30 years of research in AI, I thought that nothing could surprise me until I found a machine that knows what love is about,” said García-Crespo of the project.
The creation process, as described by García-Crespo, was similar to co-writing with another person, only at lightning speed. The AI/GPT-3 was trained using García-Crespo’s own words, then matched his writing style as the story progressed while adding new and unexpected narrative pathways. Entire storylines and paragraphs were created by GPT-3. García-Crespo praised the AI for its ability to display the sensitivity and nuances of human emotion and challenged readers to try and figure out which parts were written by the AI.
“The launching of Falta Una Palabra this coming April 23 is an invitation to readers to participate in the research study -by guessing the author of each paragraph and to immerse in the world of AI that possesses most of the attributes, skills, and characteristics to write a bestseller: a smooth and concrete storyline and coherence,” writes the AI Creative editorial team.
The novel, titled Falta Una Palabra (translated as In Need of a Word) is currently for sale on Amazon. As of today (April 25th), there are no reviews to indicate how successful the work actually is. Regardless, the work’s publication opens the door for a new frontier for AI in the arts and hints at what’s to come. I’ll be interested to see when / if AI will be able to produce entire works that pass the Turing test.
Can AI Prevent the Next Volcanic Disaster?
Shifting gears, our focus now turns to a new AI-driven method to predict volcanic eruptions. According to the article “Trust the machine: artificial intelligence may be the answer to creating rapid warning systems for volcanic eruptions,” a team of New Zealand researchers has begun to identify early warning signs of volcanic eruptions using an AI they’re calling Ardid.
“Ardid compared over 40 years of seismic data from six volcanoes, three in New Zealand and three in Alaska,” writes Cosmos’ editorial team. “Before each of [New Zealand] eruptions, a seal was forming on top of the volcano allowing pressure to build, eventually leading to an eruption. The artificial intelligence that Ardid used was able to identify when this process was happening.”
The volcanic seals indicating a potential eruption in New Zealand’s volcanoes began to show around three weeks before the event. Alaska’s volcanoes proved more unpredictable, as they never displayed a seal prior to eruption. New eruption indicators are currently being searched for by the AI.
“... the machine learning algorithm is very effective when you don’t know what patterns you’re looking for in the data, but it suspects a pattern is there,” adds the Cosmos team. According to the New Zealand team’s lead researcher, “‘It becomes a brute force technique, where it [the artificial intelligence] has this huge library of possible patterns and it just starts crunching through all of them, checking 40 years’ worth of data,’”
An early eruption warning trial system using Ardid has already been implemented at Whakaari of New Zealand’s White Island, the volcano that tragically erupted in 2019. The researchers hope to implement similar systems for other volcanic systems around the world, though it’s yet to be determined how much warning systems will be implemented. With roughly twenty-nine million people living near active volcanoes worldwide, the threat of volcanic eruption is ever-present. Look for AI to help mitigate volcanic risks in the future.
That's it for now. If you'd like to receive email updates from Aigora, including weekly video recaps of our blog activity, click on the button below to join our email list. Thanks for stopping by!