ELLIS Delft Unit shares vision on important risks of machine learning, and what would be its potential benefits

In recent months, there has been much debate about the risks of AI and machine learning (ML), a subfield of AI that has been the driving force behind many of the breakthroughs such as image recognition, playing Go, and ChatGPT. As the ELLIS Unit Delft, a collection of academics working on ML, we have faced an increasing number of questions about these risks and the debate.

Therefore, following up on the public discourse about the risks of AI, the ELLIS Unit Delft supports the statement from the ELLIS Society published on 12th June 2023. It outlines both the promises of ML as well as risks while tempering the more speculative long-term risks that could potentially overshadow the need for addressing the current concrete risks. As such it provides a balanced nuanced statement on what we see are the important risks of ML, and what would be its potential benefits.

At TU Delft, we embrace the great potential of AI to address and overcome the most pressing challenges of our era, such as health crises, inequality, the digital divide, and climate change. In the ELLIS Unit Delft, our aim is to deeply comprehend the fundamental aspects of ML as an emerging technology, specifically focusing on a number of research themes https://www.tudelft.nl/ellis-delft-unit. This aligns with our disciplinary commitment to prioritize human-centered and trustworthy technology, placing great emphasis on responsible and ethical practices.

 

ELLIS Unit Delft