Renowned Swiss organisations rely on AI- and cyber-safety startups

LatticeFlow is an Artificial Intelligence (AI) platform enabling enterprises to build performant, safe, and trustworthy AI systems at scale, helping them to unlock the full potential of AI by addressing model and data issues with automation and intelligent workflows across your AI lifecycle. The idea has gained widespread recognition and garned large projects worldwide, including one with the US army and Fortune 500 companies.

Following the launch of their new partnership in April 2023 with the aim to help businesses ensure AI safety and trustworthiness, LatticeFlow and DEKRA, a specialist in conducting independent, third-party safety and security audits for companies all over the globe, have completed their first commercial AI assessment for Migros Bank, the strategic financing partner for Switzerland's largest online car leasing provider, Gowago. At the core of Gowago's innovative solution lies an AI model that accurately predicts the residual value of cars at the end of the lease term. The assessed  AI model empowers Migros Bank and Gowago to provide competitively priced leasing terms with full transparency to their customers – which, according to Stephan Wick, COO & CIO at Migros Bank, “makes the accuracy and reliability of the AI models in use paramount in mitigating risks and ensuring regulatory compliance.”

The completion of the first commercial AI assessment coincides with the introduction of the EU AI Act and the White House mandate to establish new standards for safe, reliable, and trustworthy AI. In close alignment with these government-led initiatives, DEKRA and LatticeFlow have been collaborating with US and international government bodies, including NIST and the ISO, to offer the industry’s first AI assessment services available to enterprise clients.

xorlab supports CERN in making emails resilient to cyber attacks

For an Organization like the European Laboratory for Particle Physics, CERN, emails are one of the main forms of exchange, which regularly exposes them to cyber attacks. The potential consequences of a successful cyber attack on CERN can be enormous, considering the organization’s size.

In an effort to achieve higher resilience against cyber attacks by protecting their mailboxes against malicious, unwanted, and unsolicited emails, CERN is deploying solutions from Zurich-based startup xorlab, which builds machine-intelligent software understanding human communication behaviour and relationships to uncover even the most sophisticated email cyberattacks. Only five months after the launch, xorlab has discovered 4x more potential threats than the currently deployed solution for advanced threat protection and has stopped 19 561 phishing attacks within CERN’s emails.

To achieve these results, xorlab and the existing system processed all incoming emails of CERN for two weeks, and out of 1.9M emails received, xorlab quarantined 121K potential threats while the alternative solution detected only 30K. Upon in-depth analysis of the 102K emails quarantined by xorlab, the team found that 19% (19561) of those emails were phishing attacks and 1% (1447) other types of attacks, including BEC, extortion, malware, etc. 

(press releases / SR)