Apple Delays Approval of Email App with New ChatGPT-Style Features
Is it "unsafe for children" or just a gatekeeper asserting control?
Blix Software CEO Ben Volach has gone public with a decision by Apple to withhold approval for an update to the company’s BlueMail app due to a new GPT-3 powered feature. According to The Wall Street Journal:
“Your app includes AI-generated content but does not appear to include content filtering at this time,” Apple’s app-review team said last week in a message to the developer reviewed by the Journal.
The app-review team said that because the app could produce content not appropriate for all audiences, BlueMail should move up its age restriction to 17 and older, or include content filtering, the documents show. Mr. Volach says it has content-filtering capabilities. The app’s restriction is currently set for users 4 years old and older. Apple’s age restriction for 17 and older is for categories of apps that may include everything from offensive language to sexual content and references to drugs.
This isn’t Mr. Volach’s first tussle with the Cupertino giant. Blix Software was a founding member of The Coalition for App Fairness along with Epic Games and Spotify, which advocated for lower fees and fewer restrictions for app developers in the iOS app store. Blix also sued Apple for patent infringement, but that case was dismissed, according to 9to5Mac.
So, Volach may be a CEO with an axe to grind or merely skilled at obtaining media coverage. However, there is a bigger question here. Will the gatekeepers of Web 1 and Web 2 be able to assert control over AI adoption?
Will the gatekeepers of Web 1 and Web 2 be able to assert control over AI adoption?
Build Higher Walls
The roadblock faced by Blix may be a one-off based on unique circumstances. However, it may also be illustrative of what to expect as generative AI and synthetic media expand into more areas of daily life. Many of the winners of the Web 1 and Web 2 wars are also leaders in AI. But, right now, they don’t hold the same stranglehold in the emerging market as they do in the settled markets. That means they have the incentive to flex their existing power to steer the new technology in a way that benefits them.
When your beautiful walled garden is under assault, the immediate reaction is often to build higher walls to mitigate the impact of an unwelcome intruder. This has been the initial reaction of many educational institutions to ChatGPT. AI writing assistants are confronting some key assumptions around measuring student performance. Therefore, the best approach is to stop the advance of the technology! This is a thoroughly conservative philosophy from an education establishment that generally would position itself as the opposite of conservative.
What does Apple get out of restricting AI-generated content? I suspect very little unless it plans to introduce its own generative AI models for use by App Store developers. That is something to look for. If Apple introduces a large language model and developers using it receive preferential approval in the App Store, you will know protectionism is at play. If not, we may have a reasonable assertion of community standards or policies that should be adjusted for a new reality.
Who Will Benefit?
It suddenly seems evident to many people that AI will become very valuable. The natural reaction from those in power has been considering how they can benefit personally (or institutionally). A secondary reaction is determining how to mitigate the risk of harm. Of course, this is not primarily a concern about the risk of harm to users and everyday consumers, though that is often the core messaging. The risk that raises the most concern is anything that undermines the status quo.
Google is facing this right now. It is facing a twin assault on search and browser dominance. Grammarly is concerned that the new AI writing assistants might chip away at a market it dominated until a few weeks ago. Stock photo sites are facing an existential risk from AI-based image generators.
Ultimately, Apple might be one of the biggest near-term beneficiaries of the rise of generative AI and synthetic media. If a lot of mobile apps are able to charge monthly fees for these AI-powered services when they previously had trouble with subscription monetization, the App store may be in for a new avalanche of fees.
We should always be skeptical when institutional gatekeepers flex their power to inhibit change. Their incentives often revolve around maintaining their position. However, keep in mind that not every move by every functionary within the gatekeeping institutions of daily life is a coordinated effort to restrain or promote a particular priority.
I’m happy to give Apple the benefit of the doubt this time. Over time, the real story is typically revealed. It may be helpful now for Apple to clarify its policy. If they are saying no apps with generative text or image features should be accessible to anyone under 17, they should say for. Strategic ambiguity will lead to heightened concerns about Apple's stranglehold over mobile apps.
And do we really want to keep generative AI out of the hands of minors? Under what conditions? Let’s have the debate as opposed to just saying your app is not approved.