Anthropic does a somewhat similar thing. If you visit their ToS (the one for Max/Pro plans) from a European IP address, they replace one section with this:
Non-commercial use only. You agree not to use our Services for any commercial or business purposes and we (and our Providers) have no liability to you for any loss of profit, loss of business, business interruption, or loss of business opportunity.
It's funny that a plan called "Pro" cannot be used professionally.
Ha out of curiosity I loaded that same consumer terms URL on both a USA and a UK VPN exit node - sure enough, the UK terms inject that extra clause you quoted banning commercial usage that is not present for USA users.
There's the usual expected legal boilerplate differences. However, the UK version injects the additional clause at line 134 that has no analog in the US version.
In the Uk there seem to be separate commercial and consumer terms.
In the UK the consumer terms say its subject to English law and the courts of the UK jurisdiction you live in.
The commercial terms say that in the UK, Switzerland and the EEA there will be binding arbitration by an arbitrator in Ireland appointed by the President of the Law Society of Ireland.
The UK commercial terms explicitly do not apply to individual user plans. The US also has a separate terms sheet for commercial plans.
We are comparing like for like - an individual user using a Claude Pro subscription. A US user can use it for commercial use and be in compliance with the terms, the UK user cannot.
Will be interesting to see if those still hold any weight (in the US at least) after the latest Meta rulings established defective design as a valid reason to sue big tech for damages
show me any that have claimed that they were for entertainment purposes only. sql server has never had that in its EULA. The GPL does not say that the software is for entertainment purposes only.
Well, there's your rationale as to why AI cannot replace you.
When sh!t hits the fan, Anthropic will immediately point to this clause. Who knows, maybe a court would see it as valid.
Meanwhile, your customer (and thus, your management) is looking for someone to blame for excrement making contact with the impellers. And that someone's gonna be you.
Well, OpenAI doesn't seem to have clauses like this. Europeans are allowed to use it for commercial purposes under the ToS. (But check it yourself, I'm not a lawyer).
I reimplemented my startup idea from scratch with Codex a few months ago, just for peace of mind.
But you have limited funds to take in a lawsuit realistically the worst they can do is fire you, it's not like being blameable somehow makes you more valuable.
Employees often make mistakes that cost companies thousands of dollars. And there's no shortage of stories where employees cost companies tens of thousands and millions.
When a construction guy messes up measurements and thousands of dollars of work has the be removed and redone, no one thinks of taking the employee to court. Why would you want to take your Ai to court?
what the hell are you on about? Have you ever been employed? Employees do got reprimanded because of their mistakes. Employers just don't sue via the courts for the same reason you don't sue your spouse first thing when they break a plate. They settle via internal penalties first.
(Not only that, employees who got a reprimand too heavy handed can sue back. Plenty of cases around.)
"AI" company provides a service. They might or might not be adequate, that's not the point, the point is that the ability to sue them must always be on the cards if the agreed upon terms aren't met.
Lawyers are playing Calvinball again. I have no idea why the law finds this kind of argumentation compelling. "I clearly intentionally deceived, but I stashed some bullshit legalese into a document no one will read so my deception is completely OK."
Some 20 years ago there was a story about a guy who was opening a bank account. The bank sent the contract, the guy ameneded it with things like "you will give le unlimited credit that I do not need to repay" (if my memory serves me right).
He signed, sent both copies, got his bank signed copy back
Went yo the bank, the bank sued him, he won (the judge told the bank that when you play dirty games you sometimes loose) and they ultimately settled.
I have frequently proposed a objective legal standard for false advertising that handles that: "Technically, your honor". If somebody says that in court, they lose.
The words they used, as commonly understood by the target audience, were intentionally crafted to be interpreted differently than what they were going to say they meant in court. They spent time, effort, and money, ran focus groups, and carefully selected and curated their words to be incorrectly interpreted by the target audience to reach knowingly false conclusions.
The correct standard should be that they spent time, effort, and money, ran focus groups, and carefully selected and curated their words to be correctly interpreted by the target audience to reach true conclusions. Their statements should only be accidentally incorrect in proportion to the time and effort spent crafting and distributing them.
"Technically, your honor", should be treated as the ethical abomination it is.
On the other hand: imagine someone putting "by agreeing to this, you owe us $1,000,000,000 - unless you opt out in writing within 90 days" halfway down the 100-page EULA of some cookie-cutter smartphone app.
It is not at all uncommon for such absurd contract terms to be unenforceable - especially in B2C contracts, although it might even be tricky for B2B clickthrough ones.
The idea being that most contracts are fairly standard, so a lot of people will just skim through them. Putting a landmine in them is obviously in bad faith, so making it enforceable would basically make it impossible to do any kind of business at all.
FullStory just tried to pull this with their renewal. We had a mult-year contract that started with a two-page order form, on which the words "renewal" or "cancellation" never once appear. During negotiations, it was never discussed that the plan would renew, or that there was a cancellation window. Instead, buried at the very bottom of the form (which they send via CongaSign, and wasn't clickable or obvious), was a line about their subscription agreement being linked to their terms and conditions page. On THAT page, they mention the plan will auto renew and must be cancelled with 60 days notice.
We cancelled at T-45 or so days before renewal, having determined it wasn't a fit for our client anymore, and they insisted "well, actually, you've renewed anyway!" which, no, we haven't. Absolutely absurd to try to "clickwrap" buried renewal terms in a 20+ page T&C/privacy document rather than as a material point of fact on the actual order form being executed.
Feels like the height of absurdity to try to bully your client into forcing them to use your services against their will when they still gave ample notice that they were cancelling and when there was no material loss to the business, but it's always felt like their revenue team has been unhinged in general: exploding offers, insane terms, super high-pressure sales... part of the reason we left them in the first place.
On the other other hand, they can put whatever they want in there, and because they've forced everything into arbitration with "third party" mediation and carved out their own little niche of the justice system, they'll never actually go to court, they'll just settle and evolve their ToS and contracts and word games accordingly.
Nominally, Common Law, the system of law that to a first approximation is used in countries descended from the UK, has a lot of protections of that sort. You can't put "unconscionable" terms in a contract, e.g., it is simply illegal to sell yourself into total slavery in common-law derived systems. All signatories to a contract must consent, must not be under duress, the contract can not be one-sided (this doesn't mean "the contract is 'fair' from a 3rd-party point of view" but "the contract can't result in only one side giving things but the other doesn't"), and a variety of other common sense rules.
In practice, availing yourself of any of these protections is a massively uphill battle. Judges tend to presume that these common law matters are already embedded into the de facto legal system because the people writing the laws already operated under those assumptions while framing the law. Personally, I disagree and think a lot of these protections have eroded away into either nothing, or so little that it might as well be nothing, but you have a 0% chance of drawing me as a judge in your case so that won't help you much if you try.
I wish we lived in more of a "spirit of the law" world than a "letter of the law" world, where everything needs to be spelled out, but we don't. A small minority of people enjoy Rules Lawyering their way through life, insisting on trying to "gotcha" counterparties who are acting in good faith, so as a consequence, we all have to be Rules Lawyers and everything needs to be spelled out.
I think a “spirit of the law” world would result in judges that already abuse their absurd powers way too much have free rein over any abuse they want to do, and there would be no system for ensuring everyone is treated equally or fairly.
Theoretically, courts and judges exist precisely to balance the word and the spirit, and find and judge the actual intent. In practice, I'm in awe that good judgments still happen, despite everything.
When the contract is purposefully obtuse and hard to understand, that should be a valid legal defense.
When it's huge, falls upon people that can't justify a lawyer, and keeps changing all the time, one shouldn't even need to claim it. It should be automatically invalid.
Contract language is obtuse and hard to understand precisely because of previous challenges over meaning. There are stock phrasings and clauses in contracts that have established (by precident) legal meanings. That's why contracts seem to be walls of boilerplate.
If you just wrote them in "plain language" there would be far too much ambiguity and arguing over what was really meant or implied or agreed to.
> Copilot is for entertainment purposes only. It can make mistakes, and it may not work as intended. Don’t rely on Copilot for important advice. Use Copilot at your own risk.
Seems pretty clear to me, do you really think people need a lawyer to understand that?
Support page with ~25 tutorials provided by Microsoft about how to "Create a document with Copilot" or "Create a branded presentation from a file" or "Start a Loop workspace from a Teams meeting".
Do you actually believe that creating branded presentations (from Microsoft's own examples) is something people do for "entertainment purposes"?
If Copilot is for entertainment purposes only then why is https://office.com all about how you can use Copilot, and closes with the small print "Copilot Chat in the Microsoft 365 Copilot app is available for Microsoft 365 Enterprise, Academic, SMB, Personal and Family subscribers with a work, education, or personal account."
Why would they include a product for entertainment purposes only in the product they sell to large companies for doing work?
"Our software developers clearly were negligent, but we stashed some bullshit legalese saying 'No warranty express or implied' into a document no one will read so our bug-infested software is completely OK."
As far as I can tell, this is only for the free personal plan, not any of the business offerings (ie not Copilot for M365) and Github Copilot is under a separate set of terms.
“These Terms don’t apply to Microsoft 365 Copilot apps or services unless that specific app or service says that these Terms apply.”
Think of Copilot being a suite of different products under the same overall banner and it starts to make (a bit) more sense.
Not really since the clause in full is
"Copilot is for entertainment purposes only. It can make mistakes, and it may not work as intended. Don’t rely on Copilot for important advice. Use Copilot at your own risk."
Are you saying that the business version cannot make mistakes and can be relied upon for important advice?
I can hear the lawyers huddled around a conference table rolling the bones and chanting the sacred words to come up with that "get out of trouble free" card. It told your son he had terminal cancer and should kill himself... sorry, it clearly says for Entertainment Purposes only.
FYI: This is only for the "Cortana replacement" Copilot, not the other Copilots. This language doesn't appear in GitHub Copilot's Consumer Agreement, for example.
In the past, Microsoft named everything ".NET" [1] or "Windows Live" [2]. And before naming everything "Copilot", Microsoft named everything "Microsoft 365" [3].
Gotta love how they moved the "Create Email/meeting" buttons in Outlook mobile and stuck the Copilot button there so that you will hit it accidentally.
I’m a Mac user and the only way to get Office 365 is a monthly subscription. Since there’s no subscription that doesn’t include CoPilot and since they hiked the price with the excuse that they’d added this thing I didn’t want, I just cancelled my subscription. A customer lost: hardly an issue, but if enough people do it, maybe they’ll get a clue and stop ramming this unwelcome abomination down our throats.
I’ve only used it once, for WorkFlow creation but it seemed really useful there, but that may be more of an indictment of WorkFlow than an endorsement of CoPilot.
But according to the Birmingham modifications of 1973, subsection 12.b, stroke 7a, a player so conceding is not deemed to have actually conceded unless they be within a finite number of hops from Mornington Crescent station at the time of the concession.
I really hope so. Now I must peruse all ToS that I have agreed in the past to ensure that they have an equivalent clause. I hope I'm not contractually obliged to keep using some random website or whatever for the rest of my life.
Go to https://www.copilot.com/ and ask a question. You'll see from the answers that it is indeed for entertainment only. It is ridiculously behind ChatGPT, and I don't know how that can happen since Microsoft has access to the same models.
we can’t promise that any Copilot’s Responses won’t infringe someone else’s rights (like their copyrights, trademarks, or rights of privacy) or defame them.
You agree to indemnify us and hold us harmless (including our affiliates, employees and any other agents) from and against any claims, losses, and expenses (including attorneys' fees) arising from or relating to your use of Copilot
I thought a year ago when I bought a new laptop with 365 and Copilot integrated that they would make better use of AI and its integration. I can't think of when I actually used it and cancelled any subscription associated with it. On the otherhand, I use ChatGPT all the time.
How does this affect Copilot in VS 2022 / VS 2026? Because this is kind of insulting to a professional. I really wish Microsoft would learn to name things correctly. There's Copilot the ChatGPT-like service, then there's Copilot for Visual Studio which is not the same as far as I can tell.
> Copilot is for entertainment purposes only. It can make mistakes, and it may not work as intended. Don’t rely on Copilot for important advice. Use Copilot at your own risk.
> We don’t own Your Content, but we may use Your Content to operate Copilot and improve it. By using Copilot, you grant us permission to use Your Content, which means we can copy, distribute, transmit, publicly display, publicly perform, edit, translate, and reformat it, and we can give those same rights to others who work on our behalf.
This is as good as when the engineer from the Claude team said they load their website in such a way as to protect against hostile actions such as scraping.
Just today afternoon, I did read a bit trough Adobes EULA and I saw most of Adobes Software is not allowed to be used from children. I guess most (todays) software are not allowed for children because of the whole user tracking and spying.
Hilariously, immediately after I read this, my boss sends a global message to us reminding us that we 'need to be trying to integrate copilot into our jobs.'
It says "through other Microsoft apps and websites," i.e. they reserve the right to include or remove it when and where they like throughout their whole product line (which includes github, of course), as well as:
- Conversations you have with Copilot through third-party apps and platforms
- Other Copilot-branded apps and services that link to these Terms
That first point (#4 in the original list) can cover all software, Copilot-branded or otherwise, which, even internally, uses Copilot (perhaps without your knowing so).
Github Copilot (to take your specific example) is both "other Microsoft apps and websites" and "Copilot-branded". So, yeah, those ToS undoubtedly apply to Github Copilot.
i like the way that when ai does something good of course the people who built it should make a lot of money but when it does something bad no one is responsible
The ownership section is hilarious (tldr your content is not ours, but we can do anything you could do with it except being liable)
"We don’t own Your Content... By using Copilot, you grant us permission to use Your Content, which means we can copy, distribute, transmit, publicly display, publicly perform, edit, translate, and reformat it, and we can give those same rights to others who work on our behalf."
Anthropic does a somewhat similar thing. If you visit their ToS (the one for Max/Pro plans) from a European IP address, they replace one section with this:
Non-commercial use only. You agree not to use our Services for any commercial or business purposes and we (and our Providers) have no liability to you for any loss of profit, loss of business, business interruption, or loss of business opportunity.
It's funny that a plan called "Pro" cannot be used professionally.
https://www.anthropic.com/legal/consumer-terms
Ha out of curiosity I loaded that same consumer terms URL on both a USA and a UK VPN exit node - sure enough, the UK terms inject that extra clause you quoted banning commercial usage that is not present for USA users.
diff of the changes between US and UK:
https://www.diffchecker.com/BtqVrR9p/
There's the usual expected legal boilerplate differences. However, the UK version injects the additional clause at line 134 that has no analog in the US version.
In the Uk there seem to be separate commercial and consumer terms.
In the UK the consumer terms say its subject to English law and the courts of the UK jurisdiction you live in.
The commercial terms say that in the UK, Switzerland and the EEA there will be binding arbitration by an arbitrator in Ireland appointed by the President of the Law Society of Ireland.
The UK commercial terms explicitly do not apply to individual user plans. The US also has a separate terms sheet for commercial plans.
We are comparing like for like - an individual user using a Claude Pro subscription. A US user can use it for commercial use and be in compliance with the terms, the UK user cannot.
Software in general has disclaimed any warranties or fitness for purpose for as long as I can remember. This is nothing new.
Will be interesting to see if those still hold any weight (in the US at least) after the latest Meta rulings established defective design as a valid reason to sue big tech for damages
> Software in general has disclaimed any warranties or fitness for purpose
This is not such a disclaimer. If Copilot fails its purpose of entertaining you, you can sue. /i
Well, except that, in this case, Copilot really is for entertainment purposes only.
Prohibiting the user from using it for any commercial or business purposes is definitely new!
Have you really never seen any software saying "for non-commercial use only"?
When I'm paying for said software? No, I don't think I have.
Only when it's a free/cheap consumer version of something with a pricey business version.
show me any that have claimed that they were for entertainment purposes only. sql server has never had that in its EULA. The GPL does not say that the software is for entertainment purposes only.
Well, there's your rationale as to why AI cannot replace you.
When sh!t hits the fan, Anthropic will immediately point to this clause. Who knows, maybe a court would see it as valid.
Meanwhile, your customer (and thus, your management) is looking for someone to blame for excrement making contact with the impellers. And that someone's gonna be you.
Well, OpenAI doesn't seem to have clauses like this. Europeans are allowed to use it for commercial purposes under the ToS. (But check it yourself, I'm not a lawyer).
I reimplemented my startup idea from scratch with Codex a few months ago, just for peace of mind.
But you have limited funds to take in a lawsuit realistically the worst they can do is fire you, it's not like being blameable somehow makes you more valuable.
Employees often make mistakes that cost companies thousands of dollars. And there's no shortage of stories where employees cost companies tens of thousands and millions.
When a construction guy messes up measurements and thousands of dollars of work has the be removed and redone, no one thinks of taking the employee to court. Why would you want to take your Ai to court?
what the hell are you on about? Have you ever been employed? Employees do got reprimanded because of their mistakes. Employers just don't sue via the courts for the same reason you don't sue your spouse first thing when they break a plate. They settle via internal penalties first.
(Not only that, employees who got a reprimand too heavy handed can sue back. Plenty of cases around.)
"AI" company provides a service. They might or might not be adequate, that's not the point, the point is that the ability to sue them must always be on the cards if the agreed upon terms aren't met.
Lawyers are playing Calvinball again. I have no idea why the law finds this kind of argumentation compelling. "I clearly intentionally deceived, but I stashed some bullshit legalese into a document no one will read so my deception is completely OK."
Some 20 years ago there was a story about a guy who was opening a bank account. The bank sent the contract, the guy ameneded it with things like "you will give le unlimited credit that I do not need to repay" (if my memory serves me right).
He signed, sent both copies, got his bank signed copy back
Went yo the bank, the bank sued him, he won (the judge told the bank that when you play dirty games you sometimes loose) and they ultimately settled.
https://www.rt.com/business/man-outsmarts-banks-wins-court-2...
I can never find an article that mentions the final outcome.
I have frequently proposed a objective legal standard for false advertising that handles that: "Technically, your honor". If somebody says that in court, they lose.
The words they used, as commonly understood by the target audience, were intentionally crafted to be interpreted differently than what they were going to say they meant in court. They spent time, effort, and money, ran focus groups, and carefully selected and curated their words to be incorrectly interpreted by the target audience to reach knowingly false conclusions.
The correct standard should be that they spent time, effort, and money, ran focus groups, and carefully selected and curated their words to be correctly interpreted by the target audience to reach true conclusions. Their statements should only be accidentally incorrect in proportion to the time and effort spent crafting and distributing them.
"Technically, your honor", should be treated as the ethical abomination it is.
My two cents is that if it didn't, 'I didn't know that was illegal/breach of contract' would be a valid legal defense.
Although intentionally saying things that contradict whats in the contract might be legally objectionable.
On the other hand: imagine someone putting "by agreeing to this, you owe us $1,000,000,000 - unless you opt out in writing within 90 days" halfway down the 100-page EULA of some cookie-cutter smartphone app.
It is not at all uncommon for such absurd contract terms to be unenforceable - especially in B2C contracts, although it might even be tricky for B2B clickthrough ones.
The idea being that most contracts are fairly standard, so a lot of people will just skim through them. Putting a landmine in them is obviously in bad faith, so making it enforceable would basically make it impossible to do any kind of business at all.
FullStory just tried to pull this with their renewal. We had a mult-year contract that started with a two-page order form, on which the words "renewal" or "cancellation" never once appear. During negotiations, it was never discussed that the plan would renew, or that there was a cancellation window. Instead, buried at the very bottom of the form (which they send via CongaSign, and wasn't clickable or obvious), was a line about their subscription agreement being linked to their terms and conditions page. On THAT page, they mention the plan will auto renew and must be cancelled with 60 days notice.
We cancelled at T-45 or so days before renewal, having determined it wasn't a fit for our client anymore, and they insisted "well, actually, you've renewed anyway!" which, no, we haven't. Absolutely absurd to try to "clickwrap" buried renewal terms in a 20+ page T&C/privacy document rather than as a material point of fact on the actual order form being executed.
Feels like the height of absurdity to try to bully your client into forcing them to use your services against their will when they still gave ample notice that they were cancelling and when there was no material loss to the business, but it's always felt like their revenue team has been unhinged in general: exploding offers, insane terms, super high-pressure sales... part of the reason we left them in the first place.
On the other other hand, they can put whatever they want in there, and because they've forced everything into arbitration with "third party" mediation and carved out their own little niche of the justice system, they'll never actually go to court, they'll just settle and evolve their ToS and contracts and word games accordingly.
Not going to work in a lot of countries, again, especially with regard to consumer contracts.
Nominally, Common Law, the system of law that to a first approximation is used in countries descended from the UK, has a lot of protections of that sort. You can't put "unconscionable" terms in a contract, e.g., it is simply illegal to sell yourself into total slavery in common-law derived systems. All signatories to a contract must consent, must not be under duress, the contract can not be one-sided (this doesn't mean "the contract is 'fair' from a 3rd-party point of view" but "the contract can't result in only one side giving things but the other doesn't"), and a variety of other common sense rules.
In practice, availing yourself of any of these protections is a massively uphill battle. Judges tend to presume that these common law matters are already embedded into the de facto legal system because the people writing the laws already operated under those assumptions while framing the law. Personally, I disagree and think a lot of these protections have eroded away into either nothing, or so little that it might as well be nothing, but you have a 0% chance of drawing me as a judge in your case so that won't help you much if you try.
I wish we lived in more of a "spirit of the law" world than a "letter of the law" world, where everything needs to be spelled out, but we don't. A small minority of people enjoy Rules Lawyering their way through life, insisting on trying to "gotcha" counterparties who are acting in good faith, so as a consequence, we all have to be Rules Lawyers and everything needs to be spelled out.
I think a “spirit of the law” world would result in judges that already abuse their absurd powers way too much have free rein over any abuse they want to do, and there would be no system for ensuring everyone is treated equally or fairly.
The current alternative is the corpos abuse their absurd powers, and there's no system for ensuring everyone is treated equally or fairly.
We live in neither. Many things spelled out are unenforceable. Maybe things not spelled out are implied.
We live in a world where advertising boneless chicken does not actually mean the chicken does not contain bones.
No, you don’t. It only sounds nice. In practice this enables all kinds of spontaneous prosecution with any possible motive.
Theoretically, courts and judges exist precisely to balance the word and the spirit, and find and judge the actual intent. In practice, I'm in awe that good judgments still happen, despite everything.
When the contract is purposefully obtuse and hard to understand, that should be a valid legal defense.
When it's huge, falls upon people that can't justify a lawyer, and keeps changing all the time, one shouldn't even need to claim it. It should be automatically invalid.
Contract language is obtuse and hard to understand precisely because of previous challenges over meaning. There are stock phrasings and clauses in contracts that have established (by precident) legal meanings. That's why contracts seem to be walls of boilerplate.
If you just wrote them in "plain language" there would be far too much ambiguity and arguing over what was really meant or implied or agreed to.
> Copilot is for entertainment purposes only. It can make mistakes, and it may not work as intended. Don’t rely on Copilot for important advice. Use Copilot at your own risk.
Seems pretty clear to me, do you really think people need a lawyer to understand that?
The only thing "clear" about that License agreement is it contradicts all their other marketing about Copilot.
So either that document is fraudulent or everyone else at Microsoft is committing fraud daily.
Examples from the first search result: https://support.microsoft.com/en-us/topic/microsoft-365-copi...
Support page with ~25 tutorials provided by Microsoft about how to "Create a document with Copilot" or "Create a branded presentation from a file" or "Start a Loop workspace from a Teams meeting".
Do you actually believe that creating branded presentations (from Microsoft's own examples) is something people do for "entertainment purposes"?
Did Microsoft force you to follow the tutorials and use CoPilot for business?
If Copilot is for entertainment purposes only then why is https://office.com all about how you can use Copilot, and closes with the small print "Copilot Chat in the Microsoft 365 Copilot app is available for Microsoft 365 Enterprise, Academic, SMB, Personal and Family subscribers with a work, education, or personal account."
Why would they include a product for entertainment purposes only in the product they sell to large companies for doing work?
Microsoft is pivoting to become an entertainment company, the Copilot being the final form of what Microsoft Bob has always wanted to become.
There are 1698 words before that phrase.
Granted that this one document has a surprisingly clear language, but no, it's still not reasonable. Also, it was changed less than 6 months ago.
Sure, if you make that clear in all of your marketing rather than lying your ass off and then trying the "lol we didn’t really mean it" defense.
If it’s in a locked cabinet in the downstairs bathroom with the ‘out of order’ sign on the door, guarded by a leopard?
A disused lavatory?
We can neither confirm nor deny on advice of counsel.
"Our software developers clearly were negligent, but we stashed some bullshit legalese saying 'No warranty express or implied' into a document no one will read so our bug-infested software is completely OK."
People in glass houses shouldn't throw stones.
As far as I can tell, this is only for the free personal plan, not any of the business offerings (ie not Copilot for M365) and Github Copilot is under a separate set of terms.
“These Terms don’t apply to Microsoft 365 Copilot apps or services unless that specific app or service says that these Terms apply.”
Think of Copilot being a suite of different products under the same overall banner and it starts to make (a bit) more sense.
This should be the top comment.
Not really since the clause in full is "Copilot is for entertainment purposes only. It can make mistakes, and it may not work as intended. Don’t rely on Copilot for important advice. Use Copilot at your own risk."
Are you saying that the business version cannot make mistakes and can be relied upon for important advice?
No, I’m saying that MS have different terms for their business and personal offerings (as do OpenAI and Anthropic).
To be fair to them, MS are quite open about accuracy for the business offerings, see here as one example:
https://learn.microsoft.com/en-us/copilot/microsoft-365/micr...
I can hear the lawyers huddled around a conference table rolling the bones and chanting the sacred words to come up with that "get out of trouble free" card. It told your son he had terminal cancer and should kill himself... sorry, it clearly says for Entertainment Purposes only.
FYI: This is only for the "Cortana replacement" Copilot, not the other Copilots. This language doesn't appear in GitHub Copilot's Consumer Agreement, for example.
Maybe they shouldn't name everything Copilot.
In the past, Microsoft named everything ".NET" [1] or "Windows Live" [2]. And before naming everything "Copilot", Microsoft named everything "Microsoft 365" [3].
[1] https://en.wikipedia.org/wiki/Microsoft_.NET_strategy
[2] https://en.wikipedia.org/wiki/Windows_Live
[3] https://en.wikipedia.org/wiki/Microsoft_365
I'm old enough to remember when everything was "Active". Active Directory, ActiveX, etc.
Copilot copilot Copilot copilot copilot copilot Copilot copilot.
https://en.wikipedia.org/wiki/Buffalo_buffalo_Buffalo_buffal...
The section titled
> IMPORTANT DISCLOSURES & WARNINGS
Tells us:
> You may stop using Copilot at any time.
That's an odd thing to include in a ToS.
Like when i went to my github account to withdraw all copilot consents - which i never used anyway
just to be greeted with an email that welcomed me to copilot and the free plan. No button or link to disable the thing.
> No button or link to disable the thing.
The line i initially quoted:
> You may stop using Copilot at any time.
Was incomplete. It continues with what initially appears to be a non sequitur:
> You may stop using Copilot at any time. If you want to close your Microsoft Account, please see the Microsoft Services Agreement.
It may not be a non sequitur, but may well be the only way to "opt out" of Copilot.
I don't even have a microsoft account! anything microsoft i had (like the ancient hotmail address) was deleted years ago
I am working really hard to not start using Copilot.
And belive me, if you use any Microsoft products or services they really make it hard to avoid accidentally using the damn thing.
Including adding it to your office plan and then charging you 2x.
Gotta love how they moved the "Create Email/meeting" buttons in Outlook mobile and stuck the Copilot button there so that you will hit it accidentally.
I’m a Mac user and the only way to get Office 365 is a monthly subscription. Since there’s no subscription that doesn’t include CoPilot and since they hiked the price with the excuse that they’d added this thing I didn’t want, I just cancelled my subscription. A customer lost: hardly an issue, but if enough people do it, maybe they’ll get a clue and stop ramming this unwelcome abomination down our throats.
I’ve only used it once, for WorkFlow creation but it seemed really useful there, but that may be more of an indictment of WorkFlow than an endorsement of CoPilot.
104.3a A player can concede the game at any time.
But according to the Birmingham modifications of 1973, subsection 12.b, stroke 7a, a player so conceding is not deemed to have actually conceded unless they be within a finite number of hops from Mornington Crescent station at the time of the concession.
No, as part of the Cameron rules of 2016, concession means concession, regardless of anything else (including whether or not it’s a good idea).
Does that mean I can get an ice cream?
> > You may stop using Copilot at any time.
> That's an odd thing to include in a ToS.
Maybe it's the only Microsoft product for which that's true? (It certainly feels that way, sometimes.)
Were you able to stop using Adobe software any anytime before they got in trouble for making cancelling so hard?
I doubt it is odd, I suspect almost every ToS has something similar.
I really hope so. Now I must peruse all ToS that I have agreed in the past to ensure that they have an equivalent clause. I hope I'm not contractually obliged to keep using some random website or whatever for the rest of my life.
Go to https://www.copilot.com/ and ask a question. You'll see from the answers that it is indeed for entertainment only. It is ridiculously behind ChatGPT, and I don't know how that can happen since Microsoft has access to the same models.
it not as bad as in gpt 4.1 days, but i am wondering if it is just the system prompt or what is going on
1. April is near. Perhaps copilot allowed himself to prank us all
a blanket "entertainment only" disclaimer likely wouldn't survive scrutiny for a product actively/relentlessly marketed as a productivity tool
depends how much judges are interested in bling.
These terms too are pretty entertaning.
we can’t promise that any Copilot’s Responses won’t infringe someone else’s rights (like their copyrights, trademarks, or rights of privacy) or defame them.
You agree to indemnify us and hold us harmless (including our affiliates, employees and any other agents) from and against any claims, losses, and expenses (including attorneys' fees) arising from or relating to your use of Copilot
I've been reading Jurassic Park recently. Hammond's monologue about expensive technology only being fundable via Entertainment seems very relevant.
They should tell copilot then! It seems to disagree.
- ‚Are you for entertainment purposes only?‘
-‚Not at all — unless you want me to be. The short version: I’m not “for entertainment only.”‘
Edit: Ok I see it is legal framing to not be held liable, but can they just do that via ToS and let the tool itself promote something else?
That’s exactly what an LLM made for entertainment purposes only would say!
"Don’t use bots or scrapers"
Says the bot based on scraped data
> Other people may send similar Prompts as yours, and they could get the same, similar, or different Responses and Creations.
This is why I'm skeptical about all this AI coding thing...
If it is for entertainment purposes only, why am I not laughing when I use it?
Some people find being whipped while bound in leather to be entertaining.
You need a better sense of humour apparently.
Cool, I'm going to put this disclaimer in my work email signature. So I'm never accountable for any mistakes.
I thought a year ago when I bought a new laptop with 365 and Copilot integrated that they would make better use of AI and its integration. I can't think of when I actually used it and cancelled any subscription associated with it. On the otherhand, I use ChatGPT all the time.
How does this affect Copilot in VS 2022 / VS 2026? Because this is kind of insulting to a professional. I really wish Microsoft would learn to name things correctly. There's Copilot the ChatGPT-like service, then there's Copilot for Visual Studio which is not the same as far as I can tell.
https://docs.github.com/en/copilot/responsible-use/chat-in-y...
They do seem to word this at a more professional level in this context (the terms linked are for individuals using Copilot in Windows, probably?)
And yet my licensed Visual Studio 2026 that's supposed to be for serious-face commercial development is rife with Copilot.
> You may stop using Copilot at any time.
But how? Microsoft has shoved it into so many products that I don't see how it's possible, without dropping them alltogether.
The link is pretty clear that it is only about copilot.com, not every product that has copilot in the name:
> These Terms apply to your use of “Copilot,” which includes:
> The standalone Copilot apps on your computer or mobile device
> The Copilot service we offer at copilot.microsoft.com, copilot.com, and copilot.ai
> Conversations you have with Copilot through other Microsoft apps and websites
> Conversations you have with Copilot through third-party apps and platforms
> Other Copilot-branded apps and services that link to these Terms
> These Terms don’t apply to Microsoft 365 Copilot apps or services unless that specific app or service says that these Terms apply.
> Copilot is for entertainment purposes only. It can make mistakes, and it may not work as intended. Don’t rely on Copilot for important advice. Use Copilot at your own risk.
> We don’t own Your Content, but we may use Your Content to operate Copilot and improve it. By using Copilot, you grant us permission to use Your Content, which means we can copy, distribute, transmit, publicly display, publicly perform, edit, translate, and reformat it, and we can give those same rights to others who work on our behalf.
lol
This is as good as when the engineer from the Claude team said they load their website in such a way as to protect against hostile actions such as scraping.
Can I get this on a sticker to pass out anyone tries to shove copilot down my throat at work?
Maybe a shirt, could sell it on the Microsoft store even. Now that would be entertainment.
Just today afternoon, I did read a bit trough Adobes EULA and I saw most of Adobes Software is not allowed to be used from children. I guess most (todays) software are not allowed for children because of the whole user tracking and spying.
It could also be that minors aren't allowed to sign contracts, which a EULA could maybe be considered (I'm not a lawyer)
Hilariously, immediately after I read this, my boss sends a global message to us reminding us that we 'need to be trying to integrate copilot into our jobs.'
To us, all the profit. To you, all the risk.
+1
Software in general is usually provided on an "as is" basis with the creator not taking responsibility for anything going wrong.
To be clear this is only for the standalone Copilot chat or app and website; not for the “Copilot” services integrated into Office 365 etc.
> To be clear this is only for the standalone Copilot chat or app and website; not for the “Copilot” services integrated into Office 365 etc.
The section titled "WHEN & WHERE THESE TERMS APPLY" includes:
> Conversations you have with Copilot through other Microsoft apps and websites
Would be nice to know if it includes Github Copilot. I can't understand how to interpret "Copilot branded apps".
It says "through other Microsoft apps and websites," i.e. they reserve the right to include or remove it when and where they like throughout their whole product line (which includes github, of course), as well as:
- Conversations you have with Copilot through third-party apps and platforms
- Other Copilot-branded apps and services that link to these Terms
That first point (#4 in the original list) can cover all software, Copilot-branded or otherwise, which, even internally, uses Copilot (perhaps without your knowing so).
Github Copilot (to take your specific example) is both "other Microsoft apps and websites" and "Copilot-branded". So, yeah, those ToS undoubtedly apply to Github Copilot.
I thought the title was a joke until I actually read the thing.
Well, I'm entertained!
> Copilot may include advertising
I might be alone with this, but I don't find it very entertaining.
Worth noting that this is in the terms of use as of October 2025. This isn't "new".
Guys they're just disclaiming warranties relax
They unironically relaunch it as XBox Copilot tomorrow...
No way that holds up in court when they are marketing it for things other than entertainment.
They're just trying to pick up that Disney deal (Clippy rhymes with Mickey)
Do not taunt Happy Copilot Ball.
That one line is…doing A LOT of legal work.
It worked for Fox News
I think it's a little more nuanced than that.
https://www.allsides.com/news/2022-02-08-0800/media-industry...
https://www.snopes.com/fact-check/fox-news-entertainment-swi...
Another bingo square for that 'AI is gambling' post (https://news.ycombinator.com/item?id=47428541)
i like the way that when ai does something good of course the people who built it should make a lot of money but when it does something bad no one is responsible
Lots of that going around these days (and for many of the previous days, at least in the US)
I can't help but be reminded of Joe Pesci in Goodfellas:
"Funny how? I mean, funny like I’m a clown? I amuse you? I make you laugh? I’m here to fuckin’ amuse you? What do you mean funny, funny how?"
I've read the claim that he ad-libbed a lot of that too.
My employer does not allow me using software with entertainment function on company hardver.
Now what?! Do I have to uninstall Windows?
No, but you can’t use CoPilot any more.
One of the most toxic TOS I have ever had the misfortune of reading.
If it's for entertainment purposes only then why is it being shoved down our throats at every opportunity???
ARE YOU NOT ENTERTAINED?
I legit laughed for couple minutes, thank you for this comment.
It's not for your entertainment, silly, it's for theirs.
Mandatory Fun (TM)
So they finally admit that it's just a toy? Where does that leave all the mega-"productive" developers?
Non-exact software will be causing sleepless nights for non-exact legal writers.
I should ask it to produce an image of Satya Nadella in Maximus garb yelling "are you not entertained?!"
Ah yes, the new "for tobacco use only" of tech.
I told you so, dear LLM evangelists.
But they know better. They probably asked an LLM.
The ownership section is hilarious (tldr your content is not ours, but we can do anything you could do with it except being liable)
"We don’t own Your Content... By using Copilot, you grant us permission to use Your Content, which means we can copy, distribute, transmit, publicly display, publicly perform, edit, translate, and reformat it, and we can give those same rights to others who work on our behalf."
Seems fine to me for the consumer facing product terms lol