Examining the Intersection of Philosophy and Artificial Intelligence for Ethical Computing Insights
Striking a Balance: Ethics in AI and Computing
In the heart of MIT, Professor Armando Solar-Lezama poses a timeless query to his class, class 6.C40/24.C40 (Ethics of Computing): "How do we ensure a machine does exactly what we want, and nothing more?" This question, though seeming fresh in today's generative AI era, has been a persistent struggle since the dawn of humanity.
To illustrate this point, Solar-Lezama recounts the tale of King Midas, a monarch who received a godlike ability, but found his wish backfiring in unforeseen ways. "Be cautious about what you seek, for it may be granted in ways you don't expect," he cautions his students, many of whom aspire to become mathematicians and programmers.
Diving into the annals of MIT's archives, the professor showcases black-and-white images of the early days of programming. From the intricate cues required by the 1970s Pygmalion machine to the late '90s software that demanded teams of engineers and an 800-page manual, the methods of the past were detailed but lacked agility. They offered no room for spontaneity, play, or innovation.
Today's modern machines, while capable of incredible feats, sometimes disregard a programmer's instructions or respect nonexistent red lines. Capable of both preserving lives and inflicting harm, the consequences of these advanced machines are significant.
Titus Roesler, an electrical engineering senior, is well aware of these risks. His final paper delves into the ethical implications of autonomous vehicles, posing questions about the moral responsibility in hypothetical accidents. Roesler's argument bases itself on the philosophy theory of utilitarianism, a concept that emphasizes actions bringing the most good to the greatest number.
Listening attentively is MIT philosopher Brad Skow, who collaborates with Solar-Lezama on the course.
A Class for the Ages
Ethics of Computing, first offered in Fall 2024, is a product of the Common Ground for Computing Education, an initiative of the MIT Schwarzman College of Computing. By uniting multiple departments, the program aims to create new courses and programs that blend computing with other disciplines.
The course alternates between the lectures of Skow, a philosophy professor, who offers his discipline's unique perspective on contemporary ethical issues, and Solar-Lezama, computer science and AI lab associate director, who provides his insights. Together, they respond and adapt their lessons based on each other's classes, fostering a dynamic and engaging learning environment.
Born from Caspar Hare's desire to create an impactful, semester-long course that delved deeper than right and wrong, Ethics of Computing demands both technical and philosophical expertise. This course is a rare find at MIT, blending the two subjects in meaningful ways.
Senior Alek Westover, a mathematics and computer science double major, was drawn to the class due to his interest in AI's future trajectory. In Ethics of Computing, he has learned how to construct written arguments for complicated philosophical questions, many without a single, definitive answer.
Are We Witnessing the Demise of the World due to the Internet?
The semester unfolds with a comprehensive exploration of AI risk, examining free will, decision-making under uncertainty, long-term AI liabilities, and regulation. A second unit delves into the World Wide Web's social impact and the significance of technical decisions. The final weeks tackle privacy, bias, and free speech.
One class discussion questioned: "Are we, in essence, witnessing the internet's destruction of the world?"
Caitlin Ogoe, a senior majoring in Course 6-9 (Computation and Cognition) and a self-proclaimed technology skeptic, found the class enlightening, as it allowed her to examine pressing issues. Growing up in a family where she routinely assisted with tech support and programming, Ogoe developed a passion for compassionate technology that places people at its core. Her interest in humanities classes dealing with public policy, technology, and culture eventually led her to Ethics of Computing, marking the first course she's taken with a philosophy professor.
During a class session on bias in AI, Ogoe, a soon-to-be workplace entrant yet aspiring law school student, raises her hand to share her perspective on the controversial COMPAS AI software—an algorithm used to predict the likelihood of re-offending. Framed within the context of conflicting fairness theories, the class discusses the implications of the system's potential biases and whether it should be discontinued.
"Who knows?" ponders Solar-Lezama. "Perhaps one day the world will laugh at the concerns surrounding AI's existential risk. However, the consistent theme in this class is learning to approach these debates objectively, delving deeper than media hype, and developing a rigorous understanding of these pressing issues."
- Professor Armando Solar-Lezama, in his Ethics of Computing class, insists on finding a balance to ensure machines follow only intended commands.
- The story of King Midas teaches a lesson about unforeseen consequences of unprecedented abilities, a cautionary tale for computational fellows and professors.
- In the past, programming methods were intricate and lengthy, lacking flexibility for innovation in engineering and computing.
- Modern machines, while advanced, sometimes disregard instructions or ignore red lines, hence the significance of ethical considerations in their development.
- Titus Roesler, an electrical engineering senior, examines the ethical implications of autonomous vehicles in his final paper, using the theory of utilitarianism.
- MIT philosopher Brad Skow collaborates with Solar-Lezama to create a dynamic learning environment in the Ethics of Computing course.
- The Ethics of Computing course, offered for the first time in Fall 2024, is an initiative of the MIT Schwarzman College of Computing, blending computing with multiple disciplines.
- The course alternates lectures between Skow and Solar-Lezama, with each responding and adapting their lessons based on the other's classes.
- Senior Alek Westover, a double major in mathematics and computer science, finds the course enlightening, acknowledging the complicated philosophical questions it poses without definite answers.
- The course explores AI risk, social impact of the World Wide Web, privacy, bias, and free speech, opening up discussions about pressing issues such as the internet's potential impact on the world.
- Caitlin Ogoe, a senior and self-proclaimed technology skeptic, finds the class enlightening, fostering her passion for compassionate technology that prioritizes people.
- In a class session on bias in AI, Ogoe shares her perspective on the COMPAS AI software, prompting discussions on the implications of the system's potential biases and their resolution.
