Invisible link to canonical for Microformats
banner The Personal Press |

The Personal Press |

Card image cap

The Personal Press

In fact, just how objective is the news?

Right now, our individual choice of news platforms are already shaping our worldviews. Words essentially define the perception of a topic we are reading about. We can see that in different news outlets: some have a very conservative perspective, while others are focused on being ''social media-able''. So, what if language models enhance this phenomenon? What if every individual sees their own personalised news? What if what I see is not the same as what you see?

Generative AI can rapidly produce unique content tailored to individual readers. Companies like BuzzFeed and Meta are already making use of Language Models, to write many unique texts for many unique readers. The students behind the Coloured Realities project explore the implications of this development by creating a near-future artifact: an AI-generative newspaper. The front-page prototype serves as a tangible representation of this generative newspaper, offering a preview of a scenario where hyper-personalization of news is reality. The exhibit aims to spark discussions about the role of Language Models (like ChatGPT) being used to write news.

The visitors are encouraged to interact with the installation. Using the two dials, the reader can explore these alternate "reality bubbles", revealing how slight variations can subtly change one's perception of a subject. One dial determines your values as a reader, while the other sets the tone of the article. By turning the dial, the front-page transforms, creating an AI-generated news article that is unique to every visitor. The AI writes content based on the same key facts and arguments from different news articles on the same topic. Based on these prompts, the algorithm highlights some facts and arguments to suit the profile of the reader. These prompts wield considerable influence, raising further questions about who should be responsible for these prompts.

Card image cap
Card image cap

Card image cap

About the designers

Card image cap
This team of makers is made up of four MsC students studying Design for Interaction at TU Delft, namely:

Neslihan: a playful designer who brings her designs to life through her sketches.
Sukriti: as a designer she believes that design is a not a skillset but a mindset.
Kim: she uses design to address complex social subjects and leave positive impact.
Koen: a hands-on experimental designer who brings life to his designs by making.

As a team, they recognised that basic human rights - privacy, freedom of expression, freedom of opinion or assembly and association - are all changed when digital technology and AI is at the heart of how we communicate, socialise and interact. Freedom of expression is shaped by the platforms where we communicate. The control of information used to make decisions is invisible and complex.

Of course, generative AI was all the hype this year with lots of buzz around the question of whether it was becoming conscious. There was less focus on how it might actually change the world in the coming years. By further researching and experimenting with large language models (LLMs), and applying it to news media, the team felt like they were onto something.

The project started in a course called Interactive Technology Design. Students were invited to experiment with many different kinds of interactive technologies. At the same time, they read extensively on digital human rights – which became the topic of their case. The students looked at how they might convey a future interaction by means of a series of rapid prototypes and tests. Reflecting on these test results, the students came to understand the elements which would help them convey the message. Compiling these reflections helped them form the first prototype of a generative newspaper. Having a strong concept in their hands, it was a challenge to keep the installation simple, while maintaining the essence of the message they wanted to convey.

What we do at TU Delft
Industrial Design Engineering?

Design researchers at TU Delft work to make sure that human needs are incorporated into the rapidly developing world of AI. They collaborate with computer scientists and ethics experts in real world settings and help translate abstract AI concepts into tangible designs to generate new knowledge. A new collaborative partner in this research is - of course - AI itself. Professor Peter Lloyd’s Designing Intelligence Lab explores the creative collaboration between humans and AI. Students too are experimenting with this new technology in many courses including the Bachelor’s Digital Product Development course and a Master’s elective: Engineering with AI. Dino Liao recently graduated with a project which explored the use of AI and 3D printing technologies to enable people to design and make their own unique lamp.