Purdue University President Moon Chan made the following remarks at the university’s spring commencement ceremony held May 12-14.
opening
Today is not just a graduation, it’s a graduation ceremony in a special place, Purdue, with a rich and unique history and an accelerating momentum for excellence at scale. Nothing is more exciting than watching thousands of boilermakers celebrate the milestones of your life with those who have supported you. And this graduation has a special meaning for me as I am first in a new role serving the university.
Honorary President Mitch Daniels delivered 10 Commencement speeches, each containing an original thesis, throughout the Daniels decade. I was tempted to simply ask a generative AI engine to write this for me. But before it gets uncouth, I thought it would be more fun for humans to say a few themed words for the same humans.
Purdue AI
In the mid-20th century, AI briefly gained prominence. Now it happened again. So “hot” that any calculation seems too basic to designate itself as an AI, and any challenge seems too grand to reach. But the more we learn how tools such as machine learning work, the less the mystery fades.
For now, let’s assume that AI will eventually transform every industry and everyone, changing our lives, shaping our beliefs, and replacing our jobs. and destroy education.
We’re still playing chess after IBM’s Deep Blue beat the world champion. After the calculator, children are still taught to add numbers. Humans learn and do things not just as survival skills, but also as fun and mental training.
That doesn’t mean we don’t adapt. With the advent of calculators, elementary schools focused less on training speed in adding numbers and more on translating real-world problems into mathematical formulas. As online search became widely available, universities taught students how to properly cite online sources.
Some are considering banning AI in education. Enforcing it will be difficult. This is also unhealthy, as students will have to work in AI-powered workplaces after graduation. We would rather Purdue evolve her AI teaching and teaching with AI.
That’s why Purdue University offers multiple majors and minor degrees, fellowships, and scholarships in AI and its applications. Some come as affordable online credentials, so consider coming back to get another Purdue degree and enjoy more of your final exam.
That’s why Purdue University will explore how best to use AI to serve its students, streamlining processes and driving efficiencies to deliver a personalized experience at scale in West Lafayette. . Machines free up human time so we can do less and watch Netflix on the couch, or use the saved time to do more and create more. .
Pausing AI research is even less realistic. Especially since AI is not a separate, well-defined and well-bounded domain. All universities and companies around the world would have to stop all research involving mathematics. My PhD co-advisor, Professor Tom Cover, did his groundbreaking work on neural networks and statistics in the 1960s that he didn’t realize would later help in what he called AI. did not. We would rather Purdue advance his AI research with a nuanced understanding of the pitfalls, limitations and unintended consequences of its implementation.
That’s why Purdue just launched a campus-wide physics AI lab. Our faculty are leaders at the intersection of virtual and physical, where AI bytes meet the atoms of what we grow, create, and move, from agricultural technology to personalized medicine. Some Purdue experts are developing AI to check and contain AI through privacy-preserving cybersecurity and fake video detection.
limit and limit
As it stands, AI is good at following rules and not good at breaking them. Reinforce patterns rather than create them. We imitate what we are given and do not imagine more than the combination. Ironically, even individualization algorithms work by first grouping many individuals into a small number of “similarity classes.”
At least for now, the more artificial intelligence advances, the more amazing human intellect will be. Deep Blue vs. Kasparov, or AlphaGo vs. Lee weren’t fair comparisons. The machine he used four orders of magnitude more energy per second. Both the biological mechanisms that produce energy from food and the amount of work we do per joule must be the envy of machines. Is AI fast and energy efficient at the same time? Can it capture energy other than electricity? One day, when that happens, and when it’s combined with sensors and robotics that touch the physical world, we’re forced to question the fundamental differences between humans and machines.
Will AI ever be able to create AI? And stop the AI?
Can AI laugh, cry, and dream? Can it contain many contradictions and contradictions, like Walt Whitman?
Regardless of how “consciousness” and “soul” are defined, does AI know itself and have a soul? Huh?
Where does the life of AI begin and end in the first place? What is the identity of one AI and how can it live without dying? Will they all come together to form one life? And if an AI were to replicate a human mind and memory, would that human life last forever?
Until more architectural breakthroughs than just speeding up silicon chips and exploding data into black-box algorithms are made, these questions will remain hypothetical.
But given enough time and a matter of time, some of these questions should eventually become a reality. So what is unique to humans? What is it about artificial intelligence that is still artificial? not. Perhaps your generation has too!
freedom and rights
If boiler manufacturers had to face these issues, perhaps it would do less harm to consider an “off switch” controlled by individual citizens than a ban by some bureaucracy. May medicines be no worse than disease, and may government regulations be less fragmented and static. This is because governments have no track record of understanding rapidly changing technology, much less micromanaging it. One might argue that government access to data or mediation of algorithms is one of the most concerning uses of AI.
What we need are basic accountability guardrails such as data usage compensation, intellectual property rights, and legal liability.
Probing whether an AI engine’s output depends on its input requires a skeptical attitude. Data tends to take advantage of itself, and machines often give humans what they want to see.
We need to maintain dissent even when it is inconvenient, and avoid philosophers masquerading as AI, even when the alternatives seem inefficient.
We need entrepreneurs who invent AI systems to compete in the free market and maximize their options independently outside the big tech oligopoly. Some of them will invent ways to destroy big data.
Where, when, and how is data collected, stored, and used? Like many technologies, AI is born neutral, but prone to abuse, especially in the name of “collective good.” . AI’s most pressing and grave nightmare today is Orwell’s ‘1984’, a surveillance state suppressing rights that he was aided and abetted by AI in four minutes from that dark prophecy three centuries later. It is the misuse of AI by the political regime.
In the age of data and machines around the world, we need verifiable principles of individual rights that reflect our constitution. For example, MOTA:
- M stands for Minimalism, performing only minimal actions on data for a given purpose.
- O is Optional: Whenever possible, each person can choose to opt out.
- T: Transparency: Individuals must be notified in all cases.
- For Appeals: Companies and governments can be sued in an independent judicial system based on the rule of law if any of the above rights are violated.
My biggest fear about AI is that it will limit individual freedom. Our greatest hope for AI is that it promotes individual freedom. It’s not about increasing homogeneity, it’s about providing more choices. Freedom and liberty to choose still prevail.
Defend the rights that have survived the past centuries leaving behind other alarming headlines.
Let students develop their ability to doubt, argue and disagree.
Let a university like Purdue show you the spectacle of intellectual conflict and the struggles of critical thinking.
close
Now about commissioning an AI engine to write this speech. After finishing my own manuscript, we asked him to write a Commencement Speech for the President of Purdue on AI.
Perhaps I wasn’t intelligent enough or didn’t trust clichés on the web, but what I wrote had little overlap with what the AI did. I may be biased, but the AI version reads like his B grade high school composition, with little specificity, originality, or humor, and is a grammatically correct synthesis. It’s so perniciously common that even adding a human to the loop to build it turned out to be futile. It’s so boring that you’ll fall asleep faster than you are now.
Perhaps most commencement speeches and strategic plans sound about the same. Universities oversimplify language models. Perhaps AI can remind us to be a little less boring in what we say and think. Maybe a bot might mutter “Don’t you ChatGPT me” as he echoes in the chamber, echoing smaller and louder, all the way down to templated syntax and overused words. not. As AI becomes smarter, more interesting humans may be born.
Well, some words were duplicated between my draft and the AI draft. So, on the occasion of Purdue Spring 2023 Graduation, I have a message for you both from my “living” Bite in Chip and his Boilermaker in person. “Congratulations, and boiler up!”
