If you could talk to your past self, what would you say?

Almost nobody gets the chance to do that, but artist and scientist Michelle Huang did thanks to artificial intelligence.

In a Twitter thread, the self-described creative coder said she took more than ten years worth of childhood diaries and used them to train an AI chatbot.

READ NEXT: Kanye West and Trump allowed back on Twitter as Musk reinstates banned users

Michelle called her diaries 'fantastic, ripe data source for my experiment' that made it possible for her to 'accurately simulate what it would be like to talk to my childhood self'.

The point of the experiment, Michelle explained, was 'focused on re-engaging with and remembering the child within me' to help heal the pain of her youth.

To build the chatbot, Michelle used GPT-3, an AI language model which uses sophisticated deep learning to produce convincing, human-style chatbots.

Basically, it takes the text entered by someone as a 'prompt' and combines this with training data to produce text inspired by the prompt.

Once the chatbot was ready, Michelle said she asked it questions and got responses that 'felt eerily similar to how I think I would have responded' during that time.

Her 'younger self' chatbot also asked her questions, which Michelle said felt 'very similar to a normal texting conversation – as if I were texting my past self in real time'.

• 

  Self-replicating robots can 'build almost anything' and become bigger machines

Describing conversations with her past self as a 'healing' experience, Michelle said: "When I told her that she was loved, cared for, and safe: the words that my past self always wanted to hear[…] it felt like I was reaching into the past and giving her a giant hug, and I felt it ripple back into the present".

She added: "Overall, this was a very trippy but also strangely affirming / healing experience that I didn't realize that I had access to. Using real data from my past self allowed me to connect with her in deeper + more tangible ways than I typically have."

READ MORE:

• Demanding weekly reports from Twitter staff is 'not unreasonable', says Elon Musk
• Turn your house into Fort Knox with huge Black Friday deals on smart home gadgets
• AI 'predicts World Cup winner' after running 100,000 simulations
• £56,000 worth of drones crash into a river during Christmas lights switch-on
• 'Invisibility cloak' looks like a Christmas jumper and blocks facial recognition AI

Source: Read Full Article