Photo 8_Audrey Lorvo.jpg
Audrey Lorvo is a senior majoring in Computer Science, Economics, and Data Science and a Socially Responsible and Ethics in Computing (SERC) Scholar. Photo: MIT

AI safety research - an important field

Audrey Lorvo is currently researching AI safety, a field that aims to ensure that increasingly intelligent artificial intelligence models remain trustworthy and beneficial to humanity. The field focuses on technical challenges such as error tolerance and aligning AI with human values, while also addressing societal issues such as transparency and accountability. Experts are also concerned about the existential risks of increasingly powerful AI.

“Ensuring AI is not misused or operates contrary to human intent is increasingly important, especially as we move closer to artificial general intelligence (AGI),” Lorvo shared.

A scholar in the Social and Ethical Responsibility of Computing (SERC) program at MIT’s Schwarzman Institute for Computing, Lorvo is interested in how AI can automate AI research and development processes. “We need to do everything we can to develop AI safely,” she stressed, calling for appropriate governance frameworks and strategies to address risks.

Lorvo is committed to understanding the technical aspects of AI safety through initiatives like the AI ​​Safety Engineering Fellowship, which allows her to research ways to guide AI development in a way that benefits humanity.

"The program helped me gain a deeper understanding of the technical issues and challenges of AI safety, thereby proposing more effective AI governance strategies," she shared.

According to Lorvo, AI pioneers are pushing the technological frontier, requiring effective policy implementation to prioritize human safety without hindering research.

The value of a multidisciplinary approach

From the moment she entered MIT, Lorvo wanted to combine science and the humanities. However, the diversity of training fields made it difficult for her to choose a specific major.

“There are many ways to improve the quality of life, and MIT offers countless avenues to explore them. Professor Joshua Angrist’s econometrics classes helped me see the value of economic research, while data science and computing intrigued me because of the growing potential of AI. We can use these tools to solve our most pressing problems and overcome serious challenges,” Lorvo said.

Lorvo also expanded her research into urban planning and international development. As she delved deeper into her field, she realized that many MIT students, especially the MIT AI Alignment group, shared her views on responsible technology.

The concept of “marginal impact” describes the additional impact of an investment of time, money, or effort in a given area. According to Lorvo, this approach helps individuals decide how to invest their resources more efficiently.

In a world of limited resources, Lorvo argues, a data-driven approach can help allocate resources to areas where they are likely to generate the greatest benefit. If we want to maximize social impact, thinking about the marginal impact of career choices is important.

Student Experience at MIT

In addition to studying, Lorvo actively participates in extracurricular activities to build meaningful experiences and connect with fellow students.

"I'm fortunate to be able to balance my studies, research, club activities, and personal interests like weight training or participating in extracurricular initiatives. MIT always has a lot of interesting clubs and events to explore," Lorvo said.

These opportunities not only broadened her horizons but also helped her fine-tune her career path. Lorvo, who is fluent in French, English, Spanish, and Portuguese, also appreciated MIT's international experience programs.

Lorvo, who interned in Santiago de Chile and Paris through the MISTI program, also participated in testing a steam condensation system designed by her team as part of the D-Lab class of 2023, in partnership with the Polytechnic School of Madagascar and the NGO Tatirano. These experiences helped her gain a better understanding of how to tackle economic inequality.

After graduation, Lorvo said he wants to continue researching AI safety and governance strategies to ensure AI is deployed safely and effectively.

“Good governance is key to the successful development of AI and helping humanity take advantage of its transformative potential. We must continuously monitor the development of AI as the technology continues to evolve,” Lorvo asserted.

(According to MIT News)