I am.

Generative portrait from transcripted voice.

In a post-selfie time where machine learning algorithms convincingly turn photos into imitated masterpieces, how can computer code be used as a medium to tell stories about human identity? My latest code-based works use the Internet as an art supply store: Big and small data, cloud computing and real-time machine learning services are combined with my custom generative computer code. This process infuses with personal meaning any resulting artwork with a sort of Personal Essentialism. I borrow meaning directly from portrait sitters' personal data and I transform it into identity metaphors where meaning can be retrieved.

During april 8th and 9th, 2016 at Internet Age Media weekend I invited attendants to describe themselves, speaking to my cloud-based artistic installation. Using the Web Speech API, their voices where transcripted into text. Then, the text was transformed or complemented with literary or philosophical passages I freely associate it with through custom semantic analysis. All in near real time. The generative collage portraits contained a personal narrative texture broken by unexpected associated inspirations and random typographic accidents. See the video below and a selection of the 160 portraits created during a weekend.

This approach to turn data into meaning in the context of contemporary portraiture will be expanded in future works. So I have created an experimental platform called beyApp ("beyond appearance") that will become a sort of virtual portrait studio. Hopefully, this Art-as-a-Service project will bring me closer to my vision of generative portraits as portals to human identity.

As a first experiment on this new platform, a modified version of the I am portraits is now free and open to everyone for a limited time. Besides custom sentiment analysis and generative techniques, there is a first intent to use machine learning networks trained by me, so the resulting artwork does not imitate other artists styles, as we frequently see these days, but my very own style. Very early code and new techniques, so things can sometimes go wrong from an aesthetic point of view.

Do you want to create your portrait from your own words? Ask for an invite at www.beyapp.com. As I did with my Stardust project, it is free and open to everyone, although the invitations are limited. UPDATE 28/1/2017: The open art experiment has ended. If you want to receive invitations to the next open art experiments, please subscribe to my updates.


Creating meaning from personal data.

Creating meaning from personal data.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

Generative portrait from transcripted voice.

See related projects:

> Subscribe for updates and art experiments invitations!