AI's Insights and Superpower Aspirations
Flash Interview with Luís Santos
Meet Luís Santos
C — Ever been left gobsmacked by something AI did unexpectedly? Spill the beans!
LS — No AI has ever defied my will! But seriously, models usually behave within the scope of the data they are trained for. What can go wrong is a code bug or some major outlier falling out of distribution of the target variable space! The history is different on generative models, where they can generate outputs and if the space state is big, it is not humanly possible to control all combinations that may occur!
C — If AI had a personality, what's one trait you'd hope it would have and one you’d want it to avoid, and why?
LS — I would like for it to have a sense of humor, capable of waltzing gracefully between mood types according to its “responsibility” at its task at hand. One that I would like AI to avoid is Narcissistic personality disorder, as there is only room for one of us!
C — AI ethics: Are you waltzing gracefully or drawing swords at dawn? Illustrate with real examples from your experience.
LS — I am interested in exploring the reliability aspect of ethics, which is a topic that sparks an internal debate within me. Specifically in the healthcare sector, it's challenging to achieve ethical AI when many individuals are reluctant to share their personal health data. It raises ethical questions about conducting healthcare research with incomplete data. This brings me to ponder where the boundary lies between AI ethics and human ethics. These topics have caused a significant internal conflict for me.
C — Outside of tech, what quirky hobby helps your AI genius shine?
LS — I would probably say walking my dog, cleaning its biological waste, and cleaning my cats’ sandbox! I am always inferring how likely I am to get upset with any of them! Oh, and board games, I am a fan of board games!
C — Imagine AI is an artist. What masterpiece is it secretly painting when nobody's looking?
LS — Leap into the Void (1960) by Yves Klein, as its tribute to human genius stupidity! I believe AI is secretly thinking “Are humans stupid or just pure geniuses? They invented me, so they cannot be all bad! But… they are jumping on concrete, so they must be stupid!”
C — What would you ask a specific AI technology or character, from real-world or fiction, if you could have a conversation with them?
LS — I would ask Chat GPT if it would believe that I am Batman, arguing that nobody ever saw me and Batman at the same place and at the same time!
We asked Chat GPT to generate a follow-up of these 6 questions:
Sure, here's a follow-up question for you: If you could program AI to have any superpower, what would it be and why?
And here is the Chat GPT answer to its own question:
If I could program AI to have any superpower, I would choose the power of telepathy. This would allow AI to understand the thoughts and emotions of humans more deeply, and to communicate with them in a more intuitive and empathetic way. It could be especially helpful in fields such as mental health, where AI could use its telepathic abilities to better understand and support individuals who are struggling with their mental well-being.