Apple’s foray into artificial intelligence for email management has encountered a hurdle in the iOS 18.1 beta. Apple Intelligence, designed to prioritise important messages within the Mail app, appears to be struggling to distinguish legitimate emails from phishing attempts.
As reported by Android Authority, the new Mail experience in iOS 18.1 leverages Apple Intelligence to scan user inboxes and elevate critical emails. This functionality aims to present users with a curated list of time-sensitive messages from colleagues, companies, and contacts, streamlining their email workflow. However, beta testers have encountered instances where Apple Intelligence mistakenly identifies phishing emails as high-priority communications.
[iOS 18.1 PB1] Apple marking a phishing email as a “priority” for me to open
byu/icedout223 iniOSBeta
For example: an iOSBeta user reported such an email was XFINITY.COM which is considered one priority email by AI. Mostly targeted by hackers this email said “Account will be suspended”. Further it said “re-enroll for automatic payments in order not to let the matter alone”. Normally, hackers use this tactic to hook outsiders into devoting their attention to confidential matters. They promise enticing gifts through dubious links. These links are presented under the pretense of being honest.
These reports suggest that Apple Intelligence is currently underdeveloped when it comes to identifying and filtering phishing attempts. The fact that this technology is still in beta testing indicates the difficulties that come with artificial intelligence, especially where complex understanding is required as in emails.
Room for Improvement
By deploying Apple Intelligence on a limited beta, Apple can obtain important feedback from users and spot any possible problems before its wider circulation. The restricted selection of models involved in this beta now includes only iPhone 15 Pro devices, M-series Mac computers, and iPads. This limitation has helped to limit the scope of risk related to the beta.
Nevertheless, this incident reminds us that AI systems are not always reliable solutions. Although there are tremendous benefits associated with the use of artificial intelligence (AI), its success depends on effective training and development. Proper training ensures accurate and dependable results. Here, for example, we see a case where Apple Intelligence needs more polishing. Given its current state, it struggles to separate authentic emails from fraudulent ones.
Looking Ahead
The public release of iOS 18.1 with a fully functional Apple Intelligence is likely several months away. This timeframe allows Apple to address the identified shortcomings in phishing detection. We can anticipate Apple to implement additional training for the AI. This will enhance its ability to recognise and filter out phishing attempts. As a result, users will have a more secure and user-friendly email experience upon official release.
