Agentic AI can now use your computer
About to give a keynote in Berlin about 'AI and Assessment', and as often happens these days, I have to revamp my presentation, as while we slept, another AI release hit the streets. This time from Anthropic.AI agents are here (automation)
Things have just got more challenging for assessment but, at the same time, exciting for AI. Claude Sonnet 3.5 upgrade has just gone agentic, in a very practical way.
It can use computers the way people do. It can look at your screen, go find stuff, analyse stuff, complete a chain of tasks, move your cursor, click on buttons, type text, fill in forms. To be clear, it understands and interacts with your computer just as you are doing now. This is practical agency.
It's a bit frightening as we're starting to see glimpses of automation that are much more powerful than we thought possible. There are all sorts of startups that may ge flooded out by this move as it does what they're trying to do. But it is white collar automation that this challenges the most.
Assessment
Any way, back to assessment. The shift is in AI's ability allows it to do open-ended functions like:
• Sit tests
• Do assignments
• Do open research
Provocations
My challenge to the audience includes:
Assessment is far too ‘text’ based
Cheating has been and still is common
L&D don’t take assessment seriously
Essay setting is easy, assessment is hard
Calculators generate numbers, GenAI generates text
The idea that AI has peaked is nonsense. It is, like as rising tide drowning what were human tasks by automating processes. This is getting more profound and consequential.
No comments:
Post a Comment