Microsoft hits back at claims it slurps your Word, Excel files to train AI models
- Reference: 1732718055
- News link: https://www.theregister.co.uk/2024/11/27/microsoft_word_excel_ai/
- Source link:
The Windows giant vehemently denies the claims. A spokesperson told The Register : "In Microsoft 365 consumer and commercial applications, Microsoft does not use customer data to train large language models without your permission."
We asked Microsoft what it meant by "permission" and if the permission was opt-in or opt-out, and the IT titan has yet to respond.
[1]
Connected Experiences has long been a part of Microsoft Office. Want to do some translation? You're probably using Connected Experiences. Transcribe a recording? Again, Connected Experiences. Do some grammar checking in Word? Connected Experiences will be analyzing your content.
[2]
[3]
The spokesperson said: "The Connected Services setting is an industry standard setting that enables features that require an internet connection. Connected experiences play a significant role in enhancing productivity by integrating your content with resources available on the web. These features allow applications to provide more intelligent and personalized services."
In recent weeks, users [4]have been looking more deeply at what Microsoft is doing with all this data, and some have worried that it is being used to train the mega-corp's internal AI systems, something Microsoft says it is not.
[5]
The suggestion was [6]circulated on social media platforms over the weekend.
A look at a consumer Windows 11 machine running Microsoft 365 2410 showed that the Connected Experiences setting was checked as on by default. But did that mean the customer's content was being used to train an AI? It's unlikely but not outside the realms of possibility.
It is, however, extremely unlikely that content produced by Education and Enterprise users of Microsoft 365 would be collected in this way. After all, security policies are in place to control the Connected Experiences option if that's a concern.
[7]Security? We've heard of it: How Microsoft plans to better defend Windows
[8]Microsoft reboots Windows Recall, but users wish they could forget
[9]Microsoft shuttering dedicated licensing education, certification site
[10]Now's your chance to try Microsoft's controversial Windows Recall ... maybe
The difficulty folks face is that despite Microsoft's protestations, its [11]privacy statement (as of November 2024) does permit it to do all manner of things with the data it collects. And how does it use that data? "As part of our efforts to improve and develop our products, we may use your data to develop and train our AI models."
In August, Microsoft [12]said it would be using consumer data from Copilot, Bing, and Microsoft Start to train Copilot's generative AI models. At the time, the biz said it would allow customers to opt out and would start displaying the opt-out control in October. It also said it wouldn't be conducting training on consumer data from the European Economic Area.
[13]
Could the same apply to documents created by people in the company's flagship productivity suite? It's quite a leap from training a generative AI on what goes into Copilot to making use of Word and Excel documets under the guise of Connected Experiences. The two are very different services.
So, on the one side, Microsoft is clear: It does not use customer data to train models. On the other hand, "we may use your data to develop and train our AI models."
As for what it means by "data" in its privacy statement, the biz writes: "You provide some of this data directly, and we get some of it by collecting data about your interactions, use, and experiences with our products."
The fact that concerns are being raised indicates some users are concerned about Microsoft's AI obsession. The Windows maker must therefore maintain clarity and transparency over what is and is not going to be absorbed into the its models. ®
Updated to add at 1645 UTC, November 27
In response to our question on what Microsoft meant by "permission," if it was an opt-out or opt-in thing, and where a customer would give this permission, the company pondered for over a day before responding: "There are circumstances where enterprise customers may want or consent to our use of their data for foundation model training, such as custom model development requested by the customer."
So now you know.
Get our [14]Tech Resources
[1] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2Z0dQL4p0bT2mC0zlRIcXMgAAAE4&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Z0dQL4p0bT2mC0zlRIcXMgAAAE4&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Z0dQL4p0bT2mC0zlRIcXMgAAAE4&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[4] https://medium.com/illumination/ms-word-is-using-you-to-train-ai-86d6a4d87021
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Z0dQL4p0bT2mC0zlRIcXMgAAAE4&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[6] https://x.com/nixcraft/status/1860532716904534227
[7] https://www.theregister.com/2024/11/25/microsoft_talks_up_beefier_windows/
[8] https://www.theregister.com/2024/11/25/windows_recall_preview/
[9] https://www.theregister.com/2024/11/25/microsoft_licensing_info_site_retirement/
[10] https://www.theregister.com/2024/11/22/microsoft_recall_release/
[11] https://www.microsoft.com/en-gb/privacy/privacystatement
[12] https://www.microsoft.com/en-us/microsoft-copilot/blog/2024/08/16/transparency-and-control-in-consumer-data-use/
[13] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Z0dQL4p0bT2mC0zlRIcXMgAAAE4&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[14] https://whitepapers.theregister.com/
Microsoft's "Connected Experiences" has existed for years in Office and from reading the terms (prior to the current AI craze) "permission" means you ticked the box to get your local weather forecast displayed in Outlook. When you asked for local weather to be displayed in Outlook, or access to the thesaurus in Word, you may not have noticed that also gave Microsoft permission to do whatever they want with all of your most private data but you should have guessed really. It says it clearly on page 75.
At least it's not Disney small print.0
The only reason. Oracle's license doesn't allow them to kill you in your sleep is the delay in getting more fees from your will
> Microsoft's "Connected Experiences" has existed for years
Yes, how else do you suppose they got enough high quality data to train GPTs 1-4 ...
I also note that for "work/school accounts", " connected experiences" is mandatory, turned on via group policy (i guess probably on by default, and requires a friendly BOFH to turn it off)
It amazes me how "work/school" are assumed to be allowed to make deleterious privacy decisions on your behalf
"more intelligent and personalized services." Will it have the intelligence to adapt to my personal preferences and uninstall itself?
Yet Another Reason ...
... to consider your continued use of Microsoft Products?
I suppose now..............
I don't use the Office suit much any more.......But, having to use GPO, registery settings, and 17 different firewall outbound rules to stop the M$ sniffing and taking is something people shouldn't have to do to use a paid for product.
I don't use OutLook any more (it's thunderbird now) because of how polluted it is with crap. If M$ continues down this road I may just throw Office under the truck and leave it there until the end of time. I will miss Word and Excel but it seems this option is no longer up to the customer anymore.
Re: I suppose now..............
Libre Office is sufficiently decent for most purposes.
Re: TUT-TUT
You know that Daddy MS knows what is best for you... (sic)
Seriously,
It has to be time for people to give them the finger wherever possible. The security risks are just waiting to be exposed.
OTOH how long will it be for your fictional story that includes doing away with a major political figure results in a knock on your door at 04:00? I'm not saying do that but as of writing this, I can write a story where someone plots to kill someone else and gets stopped in the process of doing so and NOT be sent to Gitmo. But for how long after 20th Jan 2025?
MS will steal your data (just like Google etc) if you let them. Stop using MS (and Google etc). Protect your data from the thieves. You know it makes sense.
Re: I suppose now..............
try to install outlook mobile, for example, and have a look at how many different domains it connects to, they do not even try to group them up to make it look less 'massive' (or they do the grouping already and it's really a massive data gathering operation...)
Let's say that Microsoft is lying...
...and does all sorts of egregegious slurping.
Fairly easy to show that it's leaving your PC, essentially impossible to prove what they do with the data.
And say you do prove a badness, who goes to jail?
Re: Let's say that Microsoft is lying...
>Fairly easy to show that it's leaving your PC
The majority of people using Office are using Office365 so everything goes to the cloud. Ironically those users with the most sensitive data are more likely to be using Office365 vs a local copy
>And say you do prove a badness, who goes to jail?
Did you do anything hacker-ish to prove this ? Like running wireguard or a VPN or [1]Kali Linux or anything else "cyber"?
Then you do
[1] https://www.theregister.com/2020/02/14/silly_police_infosec_parental_advice_poster/
This isn't standard Microsoft/Windows bashing but, despite how much they might protest, the reasonable assumption is that all these companies are using your data until it is demonstrated otherwise. That Pokemon Go data has military applications is the latest crowning turd in that water pipe.
That said, I did look up what Connected Services entailed after the widely distributed advice to turn it off, and it seems like "and all the other stuff Word does for you", rather than AI training. I don't remember turning it on, though.
Schrödinger's data protection
It's protected, unless you take a close look at what micros~1 is actually doing
"The Connected Services setting is an industry standard setting "
A whole family of weasel word there.
Which industry? The Microsoft industry?
Whose standard? An independent open standard, a standard set by an OSI committee dominated by Microsoft acolytes or a Microsoft standard?
What's the setting? Slurp?
I can see this outside the EU on the home versions of Windows.
Would be sceptical of it happening inside the EU or on enterprise SKUs.
"You're probably using Connected Experiences"
Oh no, I'm bloody well not! LibreOffice all the way!
"Microsoft does not use customer data to train large language models without your permission."
We asked Microsoft what it meant by "permission" and if the permission was opt-in or opt-out, and the IT titan has yet to respond.
The key word is "permission". What permission, and how is it obtained? Probably by agreeing "yes" to the Windows EULA - doesn't that give them the ability to help themselves to any file on your PC, for any purpose, at any time?