Google detects AI deepfake calls: what changes for SMEs

The context: why voice spoofing has become a business problem

In recent years, phone scams have changed profoundly. Scammers no longer just call from unknown numbers. They use spoofing techniques to make trusted numbers appear on the recipient's display. And thanks to AI-based voice synthesis models, they can replicate the voice of an executive, a bank officer or a family member with impressive fidelity.

This phenomenon, known as deepfake vishing, is growing fast. According to Gartner, by 2026 30% of enterprises will consider traditional identity verification systems unreliable, precisely because of generative deepfakes. So the problem doesn't only affect private citizens: it directly hits companies' operations.

Italian B2B and retail SMEs are particularly exposed. Many authorize bank transfers, urgent orders or access to company systems with a simple phone confirmation. As a result, a convincing synthetic voice can cause significant financial damage within minutes.

What Google announced and how the system works

As reported by TechCrunch on June 2, 2026, Google has begun rolling out a fake call detection feature on Android. The system analyzes the voice and behavioral patterns of the incoming call in real time. In particular, it picks up typical signs of synthetic voices: audio artifacts, unnatural cadence, and a lack of consistent background noise.

The feature runs directly on the device, without sending audio data to Google servers. So user privacy stays protected during the analysis. When the system detects a high probability of an artificial voice, it shows a visual warning on screen during the call itself.

That said, it's important to be clear about the limits. The system doesn't block the call automatically. Its effectiveness also depends on the quality of the deepfake model the scammer uses: the most sophisticated systems may evade detection, at least early on. So think of it as an extra layer of protection, not a definitive solution.

The immediate impact on Italian SMEs

For companies with sales, admin or customer service teams, this news has concrete implications. First of all, employees who get calls from supposed executives or suppliers now have a visual warning signal. This reduces the psychological pressure typical of CEO fraud scams, where simulated urgency is part of the tactic.

Using updated Android devices at work also becomes a security requirement, not just a tech preference. SMEs that manage fleets of company smartphones should check whether their devices are compatible with this feature. Older models, in particular, may not receive the update.

By contrast, those who mainly use iOS or company VoIP systems don't benefit directly from this protection. Coverage therefore remains partial and uneven. At SHM Studio, we recommend treating this update not as a finish line, but as a sign that the tech industry is finally taking the problem seriously.

What no one is saying: the real battleground is organizational

Google's technology solves part of the problem. But the most critical component remains the human one. According to research by Harvard Business Review, most social engineering attacks succeed not because of a lack of tools, but because of a lack of clear internal procedures.

In many Italian SMEs, there's no formal protocol for verifying a caller's identity before carrying out sensitive operations. So even with a visual warning on the phone, an employee under pressure might ignore it. That's why staff training remains the most effective short-term lever.

Beyond that, it's worth reviewing your authorization processes. For example, any urgent transfer request received by phone should require confirmation on a separate channel: company email, an internal messaging app, or a callback to a verified number. Likewise, access to system credentials should never be authorized verbally.

What to do now: three priority actions for SMEs

Companies shouldn't wait for the problem to hit before acting. It helps to identify the actions with the best balance between impact and speed of implementation.

• Update company Android devices to versions compatible with the new detection feature. Check the mobile fleet update roadmap with your IT provider.
• Introduce a dual-channel verification protocol for all sensitive operations requested by phone. In particular, payments, system access and changes to supplier master data.
• Train your staff to spot the signs of a deepfake call: artificial urgency, requests for secrecy, emotional pressure. These behavioral patterns are often easier to recognize than the synthetic voice itself.

Finally, it's wise to fold these measures into a broader digital security plan. AI solutions for managing digital risk are evolving fast, and SMEs that build a solid foundation today will be better positioned in 2027-2028, when deepfake models will be even more accessible and convincing.

Outlook: where the voice security market is heading

Google isn't the only player moving in this direction. Microsoft, Apple and several enterprise security vendors are developing similar solutions. So over the next 18-24 months, it's reasonable to expect a more structured ecosystem of anti-deepfake tools built into operating systems and business communication platforms.

Still, the arms race between defenders and scammers is bound to continue. Voice synthesis models keep improving. Technological protections will have to update just as quickly. In this scenario, SMEs that build a culture of verification today, independent of the tools available, will have a structural advantage.

If you want to explore how to integrate digital security into your digital marketing strategy and communications, or understand how AI technologies are reshaping operational risks, the SHM Studio team is available for an initial consultation. You can reach us through the contact page or explore our services for Italian SMEs.

Also, if you work in B2B and manage business relationships on digital channels, it may be worth considering how to strengthen your presence on verified platforms. For example, LinkedIn campaigns let you build a traceable, authenticated communication channel with clients and partners. Likewise, a solid SEO strategy helps make your brand recognizable and hard to impersonate online.

If you'd like to dig deeper into these topics, check out our blog, where we publish up-to-date analysis on how the digital landscape is evolving for Italian businesses.