Artificial intelligence fraud prevention platform company Darwinium UK Ltd. today released updates to its Android and iOS mobile software development kits, adding continuous in-session detection of remote access scams, screen-sharing manipulation and the device tricks behind large-scale account farming.

The company is pitching the refreshed SDKs as being suited for banks, payment providers and digital businesses contending with scams that unfold after a customer has already authenticated and with mule networks that run dozens or hundreds of accounts from the same handset.

Most fraud platforms validate trust at a single moment, typically at login or payment, through device binding, an authentication step, or a challenge. Darwinium argues that approach no longer fits a threat model in which a legitimately authenticated customer can be coached or coerced into moving money minutes later, often while a remote attacker watches the screen.

The updated SDKs flag live calls running in the background on collaboration tools such as Google Meet, Microsoft Teams and Slack huddles, environments increasingly used to social-engineer victims in real time. They also separate screen-sharing signals by context, distinguishing a benign Google Cast to a television from a TeamViewer session that hands control to an attacker.

The second focus of the release is account farming, where a single device or environment is used to operate multiple identities at scale. Darwinium says the new device intelligence catches app cloning used to maintain simultaneous accounts, repackaged or modified apps surfaced through component hash checks, multiple user profiles on a single handset and mock location or GPS spoofing intended to bypass geo-restrictions and bonus rules.

The SDKs add emulator detection for iOS apps running on macOS through PlayCover, a workaround that fraud operators have been using to scale mobile attacks from desktop environments.

Banks in Southeast Asia are already running the new capabilities against mule networks, with one banking customer linking more than 10,000 mule accounts back to their underlying devices using a combination of device intelligence and forensics.

The mobile updates feed into the company’s edge-based decisioning platform, which unifies device, behavioral and identity signals across web, application programming interface and mobile traffic. Rather than running as a standalone device fingerprinting layer, the SDKs push real-time integrity data into the same risk engine that scores the rest of the customer journey.

“Most fraud platforms validate trust at a single moment, typically at login or payment, through device binding, authentication, or a step-up challenge,” said co-founder and Chief Executive Alisdair Faulkner. “But agentic-fueled mobile fraud doesn’t happen at specific moments. It happens everywhere, in a complex path of subterfuge. A customer who has legitimately authenticated can still be coached, coerced, or silently observed into moving money.”

Faulkner added that even behavioral biometrics applied as a one-time check can miss the moment a screen share begins or a remote-access tool activates mid-session.

The updated Android and iOS SDKs are available immediately to Darwinium customers.

Darwinium is a venture capital-backed startup that has raised $35 million in funding, including a round of $18 million in October 2023. Investors in the company include U.S. Venture Partners, Blackbird Ventures Pty. Ltd., Airtree Ventures Pty. Ltd. and Accomplice.

Image: SiliconANGLE/Ideogram