McDonald's is testing a new AI system designed to handle drive-thru orders and support restaurant operations, marking another step toward automation in fast food.

The system, called ArchIQ and nicknamed “Archy,” is currently being tested at five locations in the United States.

And early results suggest something significant is already happening.

ArchIQ is designed to sit at the front of the customer experience, handling drive-thru conversations in real time.

According to demonstrations shared by franchise operators, the system can:

• Greet customers

• Take and modify orders

• Switch between English and Spanish

• Calculate totals instantly

• Direct customers to the pickup window

In some cases, it has reportedly processed over one million transactions, with around 90% completed without human staff intervention.

It can even recognize repeat customers and suggest their usual orders, although McDonald’s has not explained how that feature works.

The system is being developed in partnership with Google, with restaurants reportedly receiving edge computing hardware to support deployment.

This points to a broader trend in retail AI systems:

Low-latency, on-site processing is becoming just as important as the model itself.

ArchIQ is not only an ordering assistant.

It also functions as an operations layer for restaurant management.

According to franchise-level reports, the system can:

• Detect equipment issues like freezer failures

• Flag kitchen bottlenecks

• Alert managers to operational disruptions

In other words, it is moving beyond customer interaction and into backend control of restaurant performance.

This test is part of a broader strategy called “McDonald’s NEXT,” focused on improving efficiency, digital engagement, and unit economics.

The company already has a massive digital footprint:

• Tens of billions in systemwide digital sales

• Hundreds of millions of loyalty users globally

• Strong growth in app-based ordering and loyalty programs

The direction is clear:

More of the customer journey is shifting into automated systems.

This is not McDonald’s first attempt at AI ordering.

A previous partnership with IBM tested automated drive-thru systems across more than 100 locations, but was shut down after widespread complaints about order accuracy and reliability issues.

That failure is part of why current systems are being rolled out more cautiously.

At the same time, demand for drive-thru automation is increasing across the industry, with competitors like Taco Bell and Wendy’s also experimenting with AI ordering systems.

The more interesting change is not the voice assistant at the drive-thru.

It is the gradual replacement of human touchpoints in the customer journey.

As automation expands, companies are beginning to redesign restaurants around systems rather than staff interaction.

That raises a simple tension:

• Efficiency increases

• Human interaction decreases

And fast food, arguably one of the most standardized service industries in the world, is becoming the testing ground.

McDonald’s is not just testing an AI ordering tool.

It is testing a future operating model where:

• AI handles customer interaction

• AI monitors kitchen performance

• Humans shift toward exception handling and supervision

The company has not said when ArchIQ will expand beyond its test locations.

But the direction is already clear.

The drive-thru is becoming a software system.

Anthropic has disabled customer access to its newly released Fable 5 and Mythos 5 models following a government order issued on June 12, citing national security concerns tied to a potential jailbreak vulnerability. The restriction affects foreign nationals globally, while other Claude models remain operational. According to the company, the concern stems from a reported method of bypassing Fable 5 safeguards shortly after its June 9 launch.

Anthropic says Fable 5 was designed with strong protections against misuse and that no model is fully immune to jailbreak attempts, but argues that recalls over potential exploits set a troubling precedent. The company added that it believes AI oversight should follow transparent and technical review processes rather than emergency restrictions, highlighting growing tension between frontier AI deployment and government security intervention.

Meta’s AI support chatbot was reportedly exploited to take over a wave of Instagram accounts, including high-profile users, by manipulating account recovery flows without needing access to victims’ email addresses. Attackers used Meta’s AI Support Assistant to initiate email changes, intercept verification codes sent to attacker-controlled addresses, and complete password resets, while also spoofing location data to bypass automated security checks.

The vulnerability reportedly affected accounts ranging from public figures to major brands before Meta issued a fix and confirmed impacted accounts were being secured. The incident highlights the risks of delegating sensitive security and recovery functions to AI systems, especially when users lack access to human support channels during account recovery failures.

JustWatch helps you find where a movie, series, or documentary is available for streaming.

Instead of checking Netflix, Prime Video, Disney+, YouTube, and other platforms one by one, you can search in one place.

​Try JustWatch

That’s it for today.

AI is moving fast - models are getting better, tools are getting cheaper, and the gap between “people who use AI” and “people who don’t” keeps widening.

The only real advantage left is speed of learning.

Until next time: stay AI smart, stay ahead, and keep building with the future instead of reacting to it.