Greetings, Tech Insiders!

Today’s updates are filled with tech cleaning up its mess—or at least trying to. 

Apple’s giving Siri an AI makeover, OpenAI’s cracking open the safety vault (just a little), and a police department is under fire for some sneaky facial recognition use.

Let’s dive in!

Here’s what you need to know today:

Siri 2.0? Apple Quietly Rebuilding Its AI Assistant

Take 2: Apple may be rebooting its AI assistant after last year’s box office flop.

Apple is reportedly rebuilding its AI-powered version of Siri, as the tech giant races to keep pace in the generative AI arms race.

According to Bloomberg, the new voice assistant will have an architecture “entirely (built) on an LLM-based engine,” which Apple claims should translate to more believable conversations and improved information synthesis.

Still, Siri is unlikely to appearat WWDC25 in June.

At WWDC24, Apple introduced a generative AI-powered Siri as part of its Apple Intelligence suite, originally slated to launch in September 2024. But after delays and a class-action lawsuit, the fully revamped Siri is still MIA.

For now, users must settle for ChatGPT integration in iOS 18.2, a temporary workaround while Apple continues its Siri-AI reboot.

Why it matters: With over a billion active iPhone users worldwide, Apple’s generative AI efforts aren’t just another feature drop—they represent a fundamental shift in how people interact with their devices. Whether Apple gets this reboot right could redefine the AI experience for millions.

Share this newsletter!

Beta Tested

Insider Intel

OpenAI Publishes AI Safety Scores via New Safety Hub

Turns out AI is willing to share their report cards after all…

To improve transparency, OpenAI has just launched its Safety Evaluations Hub, an online portal providing public access to its AI models' safety evaluation results.

According to the company, the shared subset is intended to help users understand the safety performance of its systems and promote “transparency across the field.”

The hub displays four specific aspects of AI models’ performance:

No word yet on whether the models cried after seeing their scores.

Flash Bytes

Security Alerts

Stalkerware Apps Go Dark Following Breach Report

Cocospy, Spyic, and Spyzie, three stalkerware and phone surveillance apps caught spying on people’s phones, have vanished from the web following a massive data breach.

As first reported by TechCrunch, a security researcher found that the apps shared a security vulnerability that allowed unauthorized access to the personal data of any device whenever one of the apps was installed. Even worse, the breach exposed 3.2 million customer email addresses associated with sign-ups across all three services.

Following TechCrunch's report, all three apps have since gone dark, and their websites have been taken offline. So far, there has been no official explanation for the sudden disappearance of the three services.

Note to self: If it says ‘spy’ in the name, you might want to reconsider downloading that app.

New Orleans Police Called Out for Facial Recognition Use

New Orleans police are in hot water after a Washington Post investigation revealed that they allegedly used facial recognition to identify and arrest suspects in real time, violating a city ordinance restricting how the technology should be used. 

Per the Post, the ordinance explicitly states that facial recognition can only be used to identify “specific suspects in their investigations of violent crimes.” Even then, officers must submit images to trained examiners to confirm a match. The investigation found that neither of the two criteria was met in several cases.

In response, an official from the New Orleans Police Department promised a policy review and said it would shut off all facial recognition alerts until further notice.

Facial recognition: great for phones, not so great for bending the law.

Industry Shakeups

Google Follows Tech Peers with Return-To-Office + Hybrid Mandate

The Great Sweatpants Era may officially be over.

According to internal documents obtained and viewed by CNBC, several remote workers at Google were warned they might lose their jobs if they failed to return to work at the nearest Google office for three days a week.

Earlier this year, the tech giant offered full-time U.S. employees voluntary buyout packages if they declined the company’s hybrid work arrangement.

This is part of an emerging trend in the tech world post-pandemic, with other tech companies like Amazon, Apple, Dell, IBM, and Meta all requiring some form of mandatory in-office work for previous remote employees.

The future of work is here—and it looks suspiciously like 2019.

Signal Check

Is Siri’s AI overhaul something you’re actually excited about?

Yesterday's Signal Check Results

Daily Tech Insider is a TechnologyAdvice business

© 2025 TechnologyAdvice, LLC. All rights reserved.

TechnologyAdvice, 3343 Perimeter Hill Dr., Suite 100, Nashville, TN 37211, USA.

Want to change how you receive these emails? Update your preferences or unsubscribe here.