Apple Delays AI-Powered Email App Update Over Child Safety Concerns

Apple reportedly delayed the approval of a BlueMail update that uses a customized version of OpenAI's GPT-3 language model due to concerns that it could generate inappropriate content for children, Reuters reports.

Last week, Apple's app-review team sent a message to BlueMail's developer, saying that the app could produce inappropriate content for all audiences and should raise its age restriction to 17 and older or include content filtering, The Wall Street Journal highlights in a report.

Update Blocked Due to AI-related Concerns

Any AI language model's output highly depends on the data its makers trained on and the prompts users gave. OpenAI's ChatGPT is trained on a massive dataset of text from the internet, which contains a variety of topics, standpoints, and language styles.

This enables systems like ChatGPT to produce diverse, creative, and human-like responses, but it also means that, if prompted, it may generate inappropriate or harmful content.

BlueMail's new AI feature uses OpenAI's latest ChatGPT chatbot to help automate email writing based on previous emails and calendar events.

Ben Volach, the co-founder of BlueMail developer Blix Inc., tells WSJ that he disagrees with Apple's decision and that the app has content-filtering capabilities, with the age restriction currently set at four years or older.

Volach argues that this request is unjustified, as Apple users can already access apps with comparable AI features and no age restrictions. Volach stated, "Apple is making it hard for us to bring innovation to our users."

AI-powered Bing is available in the iPhone App Store with the same 17-and-up age restriction as BlueMail, while Bing on the Google Play store has no age restrictions.

Moderating Generative AI

Apple's attempt to implement an age restriction to moderate content generated by a language-model-based AI indicates the tech giant is closely monitoring the risks associated with the new technology.

To maintain the privacy and security of its products, the company has stated for some time that it carefully reviews and curates the software available on the iPhone and iPad via its App Store.

Apps should not contain "content that is offensive, insensitive, upsetting, intended to disgust, in exceptionally poor taste, or just plain creepy," according to the App Store guidelines.

In the past few months, so-called generative AI has become one of the most closely observed developing technologies, with OpenAI's ChatGPT as its primary catalyst.

While AI-powered chatbots are a relatively new field, the unpredictability of early search results and conversations has already made headlines.

The dispute between Apple and BlueMail illustrates widespread doubts regarding the readiness of language-generating artificial intelligence tools such as ChatGPT for general use.

More About AI

A student in an AI ethics class recently confessed to using OpenAI's ChatGPT to collaborate on a written assignment.

A Santa Clara University professor reported that a student in an Ethics in Artificial Intelligence (AI) class used ChatGPT to generate an essay, which the student then submitted as their own.

Stay posted here at Tech Times.