Figma is adding "AI" and made all user content eligible as input. They sent an email to all users making a big benevolent show of allowing us to opt out. But
You're opted in automatically. Opting out requires action.
ONLY team/organization leaders can opt out. Do you, for work reasons say, do work on someone else's org? You're at their mercy.
"This setting doesn't apply to files in personal drafts, where AI features are always enabled"
So despite them making a big deal about protecting users, you can't really opt out at all. Consider:
If you have personal drafts, you may now have to make a choice between paying for a professional plan indefinitely and having to delete your files, if you don't want them going right into OpenAI's gullet.
Pre Adobe purchase, I hired a contractor to do some work for me in Figma. Now unless I run around making sure no early-stage work product is in their personal drafts, OpenAI gets copies.
Since the actual product doesn't work, it would be fair to look at OpenAI as nothing more or less than a big fake piece of pageantry whereby a corporation pretends to be running a business, all in order to create a social moment where every single other business in tech sends copies of all your personal files to a secretive organization owned by, among others, Sam Altman, Peter Thiel and Elon Musk.
But because of the magical social framing of "AI", nobody thinks of it that way.
That's exactly how I think of all corporate activity. We are the only product corporations are interested in. Our data, our money, our being trapped in their hellverse.
Note in case anyone is going to bring this up: Figma's help page asserts that data sent to OpenAI is "not retained" and "not to be used for model training". My position is if data is sent to a company, that company does not, in fact, have control over whether or not they retain the data. For example consider these cases:
An employee covertly copies the data
The company is served with a valid subpeona for the data
The company is compromised by nation-state intelligence or third parties
All three of these things are things that have happened in the real world, and all three can potentially happen without it constituting a violation of the privacy policy. And this is before you get into questions like whether or not I do, or should, trust OpenAI to honor their service agreements, contracts, or promises.
@mcc tbh I'm already hating this current dumpster fire of everything being online/cloud/centrally stored data, when 90% of the time they are adequate offline only
Every commercial cloud service (where a few rent an exclusive thing to the many) has that exact problem. Including Figma in this case.
Open source is the answer, because it just runs totally in your control, on your devices. (It’s so reliable that every commercial software service chooses to build their systems with open source software)
Thankfully there’s already tonnes of open source AI, and growing
@martin@mcc I've said this before: Open Source isn't anywhere near a solution in itself. Most of Facebook's code is OS. You can do every bit as much damage with OS software.
@jwcph@martin yeah, it's not obvious to me "open source" is in opposition to "the cloud". Where does your open source run? In the cloud, right? Who owns the cloud?
Are we supposed to run the OSS on our own servers? I have some VPSes on which I run some self hosted software. It's miserable. I hate it. I can't run any more software because keeping my old WordPress install barely alive takes up all the attention span I might spend on administering something I actually want to use.
There are still many good reasons for allowing someone else to provide you services around the open source software. It’s the basis of #Moodle sustainability for example.
The difference is that open source splits the software from the service, allowing you many more options.
Don’t want to run it yourself? Pick a service provider you trust.
Don’t like that service provider any more? Move to another one.
@jwcph@martin and that doesn't even require me to learn "kubernetes" or containers or whatever the heck the new standards for deploying server side software is
@mcc also OpenAI believe they can slurp literally any data on the internet regardless of license and use it in their training, so why would an agreement with another company stop them?
@mcc Absolutely. I'd add: 4. The company outsourced its data storage to "the cloud" and that company or its employees or its contractors or its subcontractors or its compromisers helped themselves to it.
Some of us think of it that way. Nobody listens to us, but rest assured, we're sitting here helplessly thinking. My favorite description of current efforts to commercialize AI is "A core of grift surrounded by a cloud of FOMO."
@mcc the first half of that sounded a lot like IBM’s Watson - not working, pageantry. If we can get past the Blackmail as a Service, there’s hope for a future when nobody cares about it anymore
@mcc let’s ban artists copying nature while we’re at it, and also painting or writing in the style of someone else. and let’s lockdown every use of art in the world except what that one artist wants. art: no more cultural, now it’s locked down. sorry, as an artist, i disagree, very much, with this take on AI. the impulse to control and restrict is not favorable to art nor artists. copyright didn’t exist thru all human history—and humanity was the richer for it. tech doesn’t change that.
@mcc I was shocked to find out I cannot opt out of the feature for my drafts... 😳 I love Figma as a product, but yeah…this might be the start of the glory days being over. 😔
Thankfully @penpot is around and keeps improving. 👌
@duncanlock It's good to know about this, but how can I trust Penpot not to do the same thing Figma did (get sold to another, large company and suddenly change corporate policies as a result)?
@nf3xn I think it (Figma) is a rather good program, but I probably would not have trusted it enough to use it if it had been owned by Adobe at the time I started using it.
I'm not familiar with Balsamiq, though I'm curious. The problem is, no matter how good the people are, what if the company gets bought like happened to Figma?
@nf3xn The thing that had me yanking my hair out this morning was realizing that certain files owned by my company had been retained by a contractor as drafts in their own account, which was normal and unobjectionable before, but NOW meant in order to AI opt out my content I had to run around getting anyone who had copies outside the "team" to delete or AI opt out their copies
@mcc I am sorry you think I am defending Adobe’s behavior. I am not; It’s abhorrent. I am cynical and suspicious of any corporation and therefore less surprised? Any way, I am sorry for you, and hope you find a way off of that platform.
@knowuh I see, thanks! Unfortunately, the problem here is I've already finished using the platform; I'm no longer actively creating files there. But I've still got files there, from the period I used it. And I don't want to delete them.
@knowuh But it's a "cloud" app, so you can't really move them off. There's no local version of the app to move them to.
This— the risk that files you created could "go away" or be data-leaked from a cloud provider years after you're done with them— is a reason I would not usually use a cloud app, especially for work purposes. In this case, the product was unique enough— and the company seemed trustworthy enough— I decided to do it anyway. But then the company got bought :(
Add comment