Dreamland
So AI is all just dream induced? What has Microsoft been smoking?
Microsoft is notifying folks that its AI services should not be taken too seriously, echoing prior service-specific disclaimers. In an update to the IT giant's Service Agreement, which takes effect on September 30, 2024, Redmond has declared that its Assistive AI isn't suitable for matters of consequence. "AI services are not …
What has Microsoft been smoking?
In 2012, Washington and Colorado were the first states to approve legal recreational use. Since then, Americans have increasingly supported legal cannabis, and legislatures have taken up the issue more often in recent years.
The timeline fits... :)
I attended a Microsoft Edu. forum on AI a few weeks ago in the UK. The CoPilot for GitHub demonstrator actually remarked that using 'please' and 'thank you' in your statement can show a marked improvement in the responses from the AI.
I am still trying to work out if this is what passes for Microsoft humour.
Fine for XBox and Clippies and lots of eye candy.
Of course, the US (and other) governments have bought into the idea that a mega-corp is trustworthy. Perhaps the mega-corp actually bought their way into these governments?
See also: "Chat GPT is bullshit": https://link.springer.com/article/10.1007/s10676-024-09775-5
"Unless explicitly permitted, you may not use web scraping, web harvesting, or web data extraction methods to extract data from the AI services"
Microsoft CEO of AI: "I think that with respect to content that is ... on the open web ... Anyone can copy it, recreate with it, reproduce with it." - https://www.theregister.com/2024/06/28/microsoft_ceo_ai/
The cognitive dissonance is strong in this one.
It isn't really cognitive dissonance. It's along the lines of saying "if you want to read a book go to the library; if you want to own a book, buy it from a bookstore instead of stealing it from the library". Although in this case it's a library of stolen books and the librarian is complaining that you're just stealing from him all the things that he put in a lot more work to steal from the original owners - proponents of AI typically thinking that "ethics" is where you live if you're a British blonde with a lisp.
"The Online Services are for entertainment purposes; the Online Services are not error-free, may not work as expected and may generate incorrect information. You should not rely on the Online Services and you should not use the Online Services for advice of any kind. Your use of the Online Services is at your own risk."
That explains very nicely how to interpret the Office 365347 offerings.
Surely, they will be TOnS(*) shopping between services when you actually have a complaint and find a rule to tell you to GFY.
(*) TOnS: Terms Of non-Service
When generating an AI, harvest and train on everything you can slurp from the net; with or without explicit consent - It's all fair use don't you know.
Now you have you AI model make claims for its use, and get people to buy into the cult (it's a cult because they have got you to believe in power of AI whilst they see it as a revenue stream)
Now you have people (possibly even paying customers) that are using the system:
[1] Change the T's & C's to prohibit people from harvesting any data / deriving understanding from your product - There is no such thing as fair use don't you know. If you pay us enough $ we might consider a single use licence.
[1] Change the T's & C's to prohibit people from using the system for anything that might be useful because if they were to do so, and something happened that doesn't fit into your cosy description of what AI can do, then you might be held liable for the consequences.
the Online Services are not error-free, may not work as expected and may generate incorrect information. You should not rely on the Online Services and you should not use the Online Services for advice of any kind. Your use of the Online Services is at your own risk."
For all that there has always been NO WARRANTY FOR THE PROGRAM, this is still fascinating to see in print.