Advertisement

Active Hackathon

You Look Like A Thing & I Love You: GPT-3 Learns To Flirt

Janelle Shane, an optics research scientist and an AI researcher, showed how GPT-3 could be used to generate pickup lines.
You Look Like A Thing & I Love You: GPT-3 Learns To Flirt

The boundary between humans and AI have been further blurred out, with GPT-3 now being introduced to the art of flirting. 

Spearheaded by Janelle Shane, an optics research scientist and an AI researcher, the project is documented on her popular science blog, named AI Weirdness, which showed how GPT-3 can be taught to generate human-like pickup lines.

THE BELAMY

Sign up for your weekly dose of what's up in emerging technology.

Since its release last summer, GPT- 3 became a revolutionary technology that utilises deep learning to create human-like text. Diving into its technicalities — the GPT-3 defines itself as an autoregressive language model that harnesses the power of deep structured learning to create human-like literature.

For example, when Shane fed Da Vinci, the fourth component of GPT-3, into the AI, a response was created that is “I love you. I don’t care if you’re a doggo in a trenchcoat”

While on the other hand, when Shane worked with the third component, Babbage, the response line was: “I Love You, I Love You, I Love You to the confines of death and disease, the legions of earth rejoices. Woe to the World”.

Weird right? Let’s find out why.

How do the gears work?

Whether or not the transformer sputters out illogical and confusing sentences or logical ones like a man who’s in love, it’s still interesting to know how the system works. 

GPT-3 is a language prediction model that deploys an algorithm that takes a piece of linguistics and transforms it into a meaningful result. This can be executed via the training analysis carried out on a body of text used to “pre-train” it. Unlike the competition, OpenAI has leveraged a vast amount of computing resources necessary for GPT-3, which allows it to further comprehend language mechanics and design.

Deploying semantic analytics, the AI understands how to build language structures such as sentences. When employing this property, it understands the structure and how each word provides meaning to the sentence. The AI primarily educates itself by unsupervised learning and creates iterations of various sentences to see which one makes sense. 

With this technology, Shane has created a squad of courtship bots that work on varying degrees of success. 

To facilitate the project, Shane provided an “article template” prompt to the four variants of GPT-3 — Ada, Babbage, Curie and Davinci. She noted that, while trying out Davinci, the “biggest and competent” of all displayed some quirky pickup lines, the smallest of all, Ada, has “completely lost the plot.”

Davinci:

“You have the most beautiful fangs I’ve ever seen.”

“Wait, this beanie hat, is it fashionable?”

Ada:

“Body Softening Pads”

“Limit to 1 purchase in stores”

“CAPE FASHION”

With higher parameters, Davinci managed to generate some readable pickup lines. However, on the other hand, Ada, being trained on fewer parameters, displayed a bunch of phrases that sounded less human-like and more bot-generated.

Dating the future

With such innovations, it’s pretty evident in the near future that humans will be courting computers.

It’s already in action, even in India, where millennials are turning towards algorithms and AI to find the correct match. One such example is Betterhalf.ai application, created by a duo of enterprising individuals. The app works on a huge variety of user data with AI’s power to figure out which two pieces of the puzzle fit best in a relationship.

Looking at the future and understanding the trends in dating via AI will be interesting. Though AI might not provide a fail-safe way to date and have a successful relationship, it significantly reduces the chances of a couple going awry.  

More Great AIM Stories

Peter Mathew
Passionate about all things media and communications. I love being a journalist, though you can see me read a book or watch a classic film in my free time.

Our Upcoming Events

Conference, Virtual
Genpact Analytics Career Day
3rd Sep

Conference, in-person (Bangalore)
Cypher 2022
21-23rd Sep

Conference, in-person (Bangalore)
Machine Learning Developers Summit (MLDS) 2023
19-20th Jan, 2023

Conference, in-person (Bangalore)
Data Engineering Summit (DES) 2023
21st Apr, 2023

Conference, in-person (Bangalore)
MachineCon 2023
23rd Jun, 2023

3 Ways to Join our Community

Discord Server

Stay Connected with a larger ecosystem of data science and ML Professionals

Telegram Channel

Discover special offers, top stories, upcoming events, and more.

Subscribe to our newsletter

Get the latest updates from AIM
MOST POPULAR

Council Post: How to Evolve with Changing Workforce

The demand for digital roles is growing rapidly, and scouting for talent is becoming more and more difficult. If organisations do not change their ways to adapt and alter their strategy, it could have a significant business impact.

All Tech Giants: On your Mark, Get Set – Slow!

In September 2021, the FTC published a report on M&As of five top companies in the US that have escaped the antitrust laws. These were Alphabet/Google, Amazon, Apple, Facebook, and Microsoft.

The Digital Transformation Journey of Vedanta

In the current digital ecosystem, the evolving technologies can be seen both as an opportunity to gain new insights as well as a disruption by others, says Vineet Jaiswal, chief digital and technology officer at Vedanta Resources Limited

BlenderBot — Public, Yet Not Too Public

As a footnote, Meta cites access will be granted to academic researchers and people affiliated to government organisations, civil society groups, academia and global industry research labs.