News EPFL

Can we convince AI to answer harmful requests?

Published:19.12.24 — New research from EPFL demonstrates that even the most recent Large Language Models (LLMs), despite undergoing safety training, remain vulnerable to simple input manipulations that can cause them to behave in unintended or harmful ways.

New cancer models could help personalize lymphoma treatments

Published:18.12.24 — Scientists at EPFL have developed "lymphomoids," a pioneering cancer model that preserves the structure and multicellular composition of lymphoma tumors in the lab. Lymphomoids offer an innovative way to test the efficacy of lymphoma treatments and better predict individual responses.

Can AI influence election outcomes?

Published:17.12.24 — Artificial intelligence (AI) may be a weapon of mass disinformation, but a recent report has demonstrated that its impact thus far has been limited

RebuiLT project shows it's possible to build differently

Published:13.12.24 — In a real-world example of how construction materials can be reused, a team of 230 EPFL students is building a community pavilion in Ecublens, near the Lausanne campus, through a low-tech, participative approach. This bold initiative entailed overcoming a number of logistical, technical and architectural challenges.

Seeing our reflection in the AI mirror

Published:12.12.24 — Attributing human traits to machines is nothing new, but with the rise of generative artificial intelligence, anthropomorphism is taking on a new dimension. This trend raises crucial philosophical and ethical issues, while redefining our relationship with technology.