Post by Ndubuisi Raphael Agua (@nura23tech)

🤖 𝗟𝗘𝗞𝗘𝗘𝗟𝗘𝗞𝗘𝗘 𝗦𝗞𝗜𝗟𝗟 𝗟𝗔𝗕

𝗘𝗣𝗜𝗦𝗢𝗗𝗘 𝟬𝟬𝟲 — 𝗖𝗔𝗡 𝗛𝗨𝗠𝗔𝗡 𝗕𝗜𝗔𝗦 𝗔𝗙𝗙𝗘𝗖𝗧 𝗔𝗜? 🧠🤖

Good evening tribe. I trust we all had a good day and are settling in nicely for the night

Yesterday, we looked at how humans help AI learn by preparing data, labelling examples, correcting mistakes and giving feedback, but there is something else we need to understand — humans can also influence what AI learns through the choices they make 👀

Imagine we give an AI thousands of examples, but those examples mostly represent one group of people or one particular way of thinking, the AI may learn that pattern and later produce results that do not work fairly for everyone

This is where bias comes in

Bias simply means that something is not completely balanced or neutral, and it can enter AI through the data, the way examples are labelled, the questions we choose to ask, the people involved in building the system and even the way the AI is designed or used

For example, if an AI is trained mostly with information from one environment, it may not perform as well when used in a completely different environment

𝗜𝗡 𝗦𝗜𝗠𝗣𝗟𝗘 𝗧𝗘𝗥𝗠𝗦:

𝗛𝗨𝗠𝗔𝗡 𝗖𝗛𝗢𝗜𝗖𝗘𝗦 → 𝗗𝗔𝗧𝗔 → 𝗔𝗜 𝗟𝗘𝗔𝗥𝗡𝗦 → 𝗣𝗢𝗦𝗦𝗜𝗕𝗟𝗘 𝗕𝗜𝗔𝗦

This is why saying “AI said it” should never automatically mean “it must be right”

AI can be powerful, but humans still have a responsibility to question, test, check, and improve the systems we build

And that takes us to a very important question 👀

𝗜𝗙 𝗛𝗨𝗠𝗔𝗡𝗦 𝗖𝗔𝗡 𝗜𝗡𝗙𝗟𝗨𝗘𝗡𝗖𝗘 𝗔𝗜, 𝗛𝗢𝗪 𝗗𝗢 𝗪𝗘 𝗠𝗔𝗞𝗘 𝗦𝗨𝗥𝗘 𝗔𝗜 𝗜𝗦 𝗨𝗦𝗘𝗗 𝗥𝗘𝗦𝗣𝗢𝗡𝗦𝗜𝗕𝗟𝗬? 👀

That’s where AI Governance comes in

𝗟𝗲𝗮𝗿𝗻 • 𝗕𝘂𝗶𝗹𝗱 • 𝗚𝗿𝗼𝘄 • 𝗘𝗮𝗿𝗻

#LekeeLekee
#SkillLab
#Nura23Tech
#Trending

0 likes · 0 comments · 0 shares