Human-centered AI (HCAI) is an evolution of user-centered design with a core belief that AI should be used to enhance and empower people not replace them.
A brief history
Since the 1950's many AI Researchers, Scientists and Companies are building AI with the core goal of being able to replace humans with AGI.
They think that humans and computers are of the same kind and category, however, HCAI advocates know humans are not computers.
Computers are not human.
and never will be.
Like Ben Shneiderman states in Human-Centered AI “Humans have bodies. Having a body makes you human. It puts us in touch with pain and pleasure, with sadness and joy. Crying and laughing, dancing and eating, love-making and thinking are all parts of being human.”
"Computers don't contain 'brains' any more than stereos contain musical instruments... Machines only manipulate numbers; people connect them to meaning" - Arno Penzias, Nobel Prize in Physics (1989)
What is the point of emulating what Humans can already do? It is more interesting to use AI to create new tools that amplify and extend Human abilities just like:
- Exoskeletons that help paralysed or weak individuals to become fully mobile again.
- Digital cameras that enable users to quickly shoot photos that are correctly exposed and focused.
- Tele-operated Surgical robots that help surgeons save lives more effectively anywhere in the world.
- Navigation apps like Google Maps that use AI to find the fastest and most efficient route.
HCAI builds upon user-centered design thinking and methodologies to help founders, designers and product professionals implement AI that is reliable, safe and trustworthy while enabling high levels of autonomy and human control. (Yes, you can have both).
Safe, Reliable and Trustworthy
As AI develops at a rapid speed it is fundamental that it is used in a responsible way. One of HCAI’s core goals is to develop systems that are safety first, reliable and trustworthy which can be acheived with a range of methodologies (see here).
Summary of HCAI
HCAI thinking is a sure way to develop products that are commercially successful by solving human needs with autonomous systems that provide novel and creative abilities with the human at the heart.
HCAI facilitates using AI to enhance, amplify and extend human capabilities so we can do things we couldn't before and be more creative and capable than ever before.
HCAI facilitates the ability to have high levels of autonomy (AI) and high levels of human control which means we develop tools that are trusting, reliable and safe.
HCAI is pro human and pro AI playing to both their strengths but understanding they are different and will never be the same.
Human-in-the-loop is a term often said which eludes to a relationship where AI is in control and occasionally needs to confer with humans. A phrase inline with HCAI thinking is "Humans in the group, computers in the loop" coined by Ben Shneiderman.
HCAI systems implement strong safety precautions like kill switches, e2e testing and advanced logging to protect people from faulty AI implementation.