Align AI with human value | MIT News

Senior Audrey Lorvo is working on AI security, which is designed to ensure increasingly intelligent AI models are reliable and can benefit humans. The growing field focuses on technical challenges such as robustness and AI consistency with human values, as well as social issues such as transparency and accountability. Practitioners are also focusing on the potential risks associated with increasingly powerful AI tools.

“It is becoming increasingly important to ensure that AI does not abuse or violate our intentions,” said Lorvo, a major in computer science, economics and data science. AGI describes the potential of artificial intelligence to match or surpass human cognitive abilities.

MIT Schwarzman School of Computer and Ethical Responsibility (SERC), Lorvo, a scholar at MIT Schwarzman School of Computer and Ethical Responsibility, has carefully studied how AI automates the development process and development process and practice of AI. She is a member of the Big Data Research Group and is investigating the social and economic impacts associated with AI’s potential to accelerate its own research and how to effectively communicate these ideas and potential impacts to a general audience, including legislators, strategic advisers and others.

Lorvo stressed the need to critically evaluate the rapid advancement of AI and its implications to ensure that organizations have the appropriate frameworks and strategies to address risks. “We both need to ensure that humans have access to the benefits of AI and that we don’t lose control of technology,” she said. “We need to develop it safely whenever possible.”

Her efforts to participate in AI Security Technology Scholarships reflect her investment in understanding AI Security Technology. The scholarship provides the opportunity to review existing research to enable AI development with potential human impact considerations. “Scholarships helped me understand the technical issues and challenges of AI security, so I could come up with better AI governance strategies,” she said. According to Lorvo, companies in the AI ​​space continue to push boundaries, which means we need to implement effective policies to prioritize human security without hindering research.

The value of human participation

Upon arriving at MIT, Lorvo knew she wanted to take a study course that would allow her to work at the intersection of science and humanities. However, the various products from the institute made her choice difficult.

“There are many ways to help improve the quality of life for individuals and communities, and MIT offers many different avenues of investigation,” she said.

Starting with economics – she likes discipline because it focuses on quantitative impact – Lorvo investigated mathematics, political science, and urban planning before choosing courses 6-14.

“Professor Joshua’s econometrics course helped me see the value of focusing on economics, while the data science and computer science elements attracted me as the impact of AI is growing and potential,” she said. “We can use these tools to solve some of the world’s most pressing problems and hope to overcome serious challenges.”

Lorvo is also focusing on urban research, planning and international development.

As she narrowed her focus, Lorvo discovered that she shared human prospects with other members of the MIT community, such as the MIT AI Alignment Group, from which she learned a lot about AI security. “Students care about their marginal impact,” she said.

Marginal impact is other effects of a specific investment in time, money, or effort, and is a way to measure the increase in contribution rather than focusing on the total impact. This could affect where people choose to invest resources, and the idea that Lorvo is attracted.

“In a world of limited resources, data-driven approaches to address some of our biggest challenges can benefit from tailored approaches that direct people to where they may do the best,” she said. “If you want to maximize your social impact, it can be very valuable to reflect on the marginal impact of your career choices.”

Lorvo also values ​​MIT focusing on educating students as a whole and takes advantage of the opportunity to investigate disciplines such as philosophy through MIT Concourse, a program that promotes dialogue between science and humanities. The conference expects participants to receive guidance, clarity and purpose in science, technology and human pursuits.

Student experience at the Institute

Lorvo devotes his time outside the classroom to create memorable experiences and promote relationships with classmates. “I’m lucky to have the space to balance my courses, research and club commitments with other activities such as weightlifting and off-campus initiatives,” she said. “There are always a lot of clubs and activities throughout the institute.”

These opportunities to expand her worldview challenge her beliefs and expose her to new areas of interest that transformed her life and career choices. Fluent French, English, Spanish and Portuguese Lorvo also applauds MIT for the international experience provided to students.

“I have been internships in Santiago de Chile and Paris and have worked with Misti to test the water vapor condensation chamber we worked with the Madagascar Polytechnic and Tatirano Ngo (organized by the Nongovernovernmovernnal) in the fall 2023 D-LAB class, “and enjoyed the opportunity to address economic development through my scope of economic development.”

As president of the MIT Undergraduate Economics Association, Lorvo connects with other students interested in economics while continuing to expand his understanding of the field. She loves the relationship she has built and also participates in the association’s activities throughout the year. “Even as a senior, I discovered new campus community exploration and appreciation,” she said. “I encourage other students to continue exploring the groups and courses that sparked their interest during MIT.”

After graduation, Lorvo hopes to continue investigating AI security and research governance strategies to help ensure the safe and effective deployment of AI.

“Good governance is crucial to the successful development of AI and ensure that humans can benefit from its change potential,” she said. “As technology continues to evolve, we must continue to monitor the growth and capabilities of AI.”

Understand the potential impact of technology on humanity, do a good job, improve continuously, and create space where big ideas can see the light of the day continue to push Lorvo. Combining the humanities and science into most of her work. “I have always wanted to contribute to improving people’s lives, and AI represents the greatest challenge and opportunity for humanity,” she said. “I believe the field of AI security can benefit from people with interdisciplinary experience, such as the kind I’m lucky enough to have, and I encourage anyone to be keen on shaping the future to explore it.”