Lattica secures $3.25 million to boost private AI with FHE
The funding round was led by Cyber Fund, the investment firm of Managing Partner Konstantin Lomashuk, with additional participation from angel investor Sandeep Nailwal, who is a co-founder of both Polygon Network and Sentient: The Open AGI Foundation.
Lattica's technology aims to address persistent privacy and security challenges in sectors such as healthcare, finance, and government, where organisations are hesitant to adopt AI over concerns about sensitive data exposure. Referencing the Cisco 2025 AI Briefing: CEO Edition, the company highlighted that 70% of surveyed CEOs are worried about their network security due to increasing AI use, and 34% see security as a major obstacle to AI adoption.
FHE, which allows encrypted data to be queried by AI models without decryption, has long been considered a desired goal in cryptography, but has until now suffered from computational inefficiencies. Lattica claims it has managed to operationalise FHE by leveraging the latest advancements in the AI acceleration stack, making use of acceleration techniques that improve FHE's commercial viability.
Dr. Rotem Tsabary, Founder and Chief Executive Officer at Lattica, holds a PhD in lattice-based cryptography from the Weizmann Institute of Science and leads a team focused on the mathematical similarities between FHE and machine learning to build a cloud-based, hardware-agnostic platform for private AI computation.
Central to Lattica's solution is the Homomorphic Encryption Abstraction Layer (HEAL), a cloud-based service designed to enhance FHE's performance and streamline its adoption. HEAL operates as an interface connecting FHE-equipped applications with a variety of computing hardware, including GPUs, TPUs, CPUs, and purpose-built accelerators such as ASICs and FPGAs.
"By combining the advancements of hardware acceleration with software-based optimisation, we realised that not only could we improve FHE efficiency to the point of commercial viability, but use it to solve critical data dilemmas holding back AI's adoption in sensitive industries," said Dr. Rotem Tsabary, founder and CEO of Lattica. "We're enabling practical FHE by developing a solution that is tailor made for neural networks."
In conjunction with its debut, Lattica has posted demonstration versions of its platform and shared the results of an in-depth survey conducted within the FHE community. According to these findings, 71% of participants believe FHE adoption will likely depend on a hybrid approach consisting of both hardware and software innovations.
Konstantin Lomashuk, Managing Partner at Cyber Fund, commented, "Lattica is pushing the boundaries of Fully Homomorphic Encryption, solving one of the most critical challenges in AI security. Cyber Fund is proud to have led Lattica's pre-seed round. This is the kind of deep-tech innovation that defines the future, and we're excited to see Lattica leading the way."
The healthcare and finance industries are specific targets for Lattica, given the demand for secure, cloud-based AI applications that can process medical and financial data without exposing such information to providers or third-party platforms. The company sees use cases in encrypted financial transactions and secure data analysis for medical research.
Sandeep Nailwal, co-founder of Polygon Network and investor in Lattica, said, "Lattica's product-first approach fundamentally transforms sensitive data processing in the AI ecosystem. Lattica has made FHE a reality that is both practical and scalable, as Tsabary and her research team is proving that advances in the machine learning stack can significantly boost the performance of FHE and have an immediate impact on the market."
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
NZ Herald
2 days ago
- NZ Herald
ChatGPT got a big upgrade. Here's what to know about OpenAI's GPT-5
How much does it cost to access GPT-5? All ChatGPT users will get access to GPT-5, even those using the free version. But only those with a US$200-a-month ($335) 'Pro' subscription get unlimited access to the newly released system. GPT-5 will be the default mode on all versions. Users not paying for ChatGPT will only be able to ask a certain number of questions answered by GPT-5 before the chatbot switches back to using an older version of OpenAI's technology. How will GPT-5 change ChatGPT? GPT-5 responds to questions faster than OpenAI's previous offerings and is less likely to 'hallucinate' or make up false answers, OpenAI executives said at a news briefing before its release. It gives ChatGPT 'better taste' when generating writing, said Nick Turley, who leads work on the chatbot. OpenAI's new AI software can also answer queries using a process dubbed reasoning that shows the user a series of messages attempting to break down a question into steps before giving its final answer. 'GPT-5 is the first time that it really feels like talking to an expert, a PhD-level expert,' OpenAI CEO Sam Altman said. Altman said GPT-5 is particularly good at generating computer programming code, a feature that has become a major selling point for OpenAI and rival AI developers and has transformed the work of programmers. In a demo, the company showed how two paragraphs of instruction was enough to have GPT-5 create a simple website offering tutoring in French, complete with a word game and daily vocabulary tests. Execs say ChatGPT users can now connect the app with their Google calendars and email accounts. Photo / Getty Images Altman predicted that people without any computer science training will one day be able to quickly and easily generate any kind of software they need to help them at work or with other tasks. 'This idea of software on demand will be a defining part of the new GPT-5 era,' Altman said. Turley also claimed the upgrade made ChatGPT better at connecting with people. 'The thing that's really hard to put into words or quantify is the fact that just feels more human,' he said. In a livestream Thursday, OpenAI execs said ChatGPT users could now connect the app with their Google calendars and email accounts, allowing the chatbot to help people schedule activities around their existing plans. What does it mean for an AI chatbot to 'reason?' GPT-5 could give many people their first encounter with AI systems that attempt to work through a user's request step-by-step before giving a final answer. That so-called 'reasoning' process has become popular with AI companies because it can result in better answers on complex questions, particularly on math and coding tasks. Watching a chatbot generate a series of messages that read like an internal monologue can be alluring, but AI experts warn users not to confuse the technique with a peek into AI's black box. The self-chatter doesn't necessarily reflect an internal process like that of a human working on a problem, but designing chatbots to create what are sometimes dubbed 'chains of thought' forces the software to allocate more time and energy to a query. OpenAI released its first reasoning model in September for its paying users, but Chinese start-up DeepSeek in January released a free chatbot that made its 'chain of thought' visible to users, shocking Silicon Valley and temporarily tanking American tech stocks. The company said ChatGPT will now automatically send some queries to the 'reasoning' version of GPT-5, depending on the type of conversation and complexity of the questions asked. Is GPT-5 the 'super intelligence' or 'artificial general intelligence' OpenAI has promised? No. Tech leaders have for years been making claims that AI is improving so fast it will soon become able to learn and perform all tasks that humans can at or better than our own ability. But GPT-5 does not perform at that level. Super intelligence and artificial general intelligence, or AGI, remain ill-defined concepts because human intelligence is very different from the capabilities of computers, making comparisons tricky. OpenAI CEO Altman has been one of the biggest proponents of the idea that AI capabilities are increasing so rapidly that they will soon revolutionise many aspects of society. 'This is a significant step forward,' Altman said of GPT-5. 'I would say it's a significant fraction of the way to something very AGI-like.' Some people have alleged that loved ones were driven to violence, delusion or psychosis by hours spent talking to ChatGPT. Photo / Getty Images Does GPT-5 change ChatGPT's personality? Changes OpenAI made to ChatGPT in April triggered backlash online after examples of the chatbot appearing to flatter or manipulate users went viral. The company undid the update, saying an attempt to enhance the chatbot's personality and make it more personalised instead led it to reinforce user beliefs in potentially dangerous ways, a phenomenon the industry calls 'sycophancy'. OpenAI said it worked to reduce that tendency further in GPT-5. As AI companies compete to keep users engaged with their chatbots, they could make them compelling in potentially harmful ways, similar to social media feeds, The Washington Post reported in May. In recent months, some people have alleged that loved ones were driven to violence, delusion or psychosis by hours spent talking to ChatGPT. Lawsuits against other AI developers claim their chatbots contributed to incidents of self-harm and suicide by teens. OpenAI released a report on GPT-5's capabilities and limits Thursday that said the company looked closely at the risks of psychosocial harms and worked with Microsoft to probe the new AI system. It said the reasoning version of GPT-5 could still 'be improved on detecting and responding to some specific situations where someone appears to be experiencing mental or emotional distress'. Earlier this week, OpenAI said in a blog post it was working with physicians across more than 30 countries, including psychiatrists and paediatricians, to improve how ChatGPT responds to people in moments of distress. Turley, the head of ChatGPT, said the company is not optimising ChatGPT for engagement.
Techday NZ
2 days ago
- Techday NZ
Exclusive: Garrett O'Hara on Mimecast's AI fight against cyber risk
In a world where cyberattacks are growing more sophisticated and frequent, organisations are increasingly focusing on what Garrett O'Hara calls the "most unpredictable element in security" - humans. Speaking during a recent interview, Garrett O'Hara, Senior Director of Solutions Engineering for APAC at Mimecast, explained how artificial intelligence (AI) is now being deployed to manage and mitigate human risk at scale. "Human risk is anything people can do that exposes an organisation to risk, either by accident or intent," he said. "Most of the time, it's not malicious - it's tiredness, deadlines, or someone trying to do their job more efficiently." He pointed out that employees often unintentionally bypass security policies under pressure. "They might upload sensitive documents to a personal drive just so they can work from home, not realising the huge risk that introduces," he added. AI tools, while offering productivity benefits, have also opened new doors for attackers. "We're seeing employees use tools like ChatGPT to summarise documents or create presentations, not realising they're potentially uploading sensitive corporate data to third-party platforms," he said. On the flip side, O'Hara said AI is a vital asset in the fight against these new types of threats. "AI is incredibly good at detecting patterns and threats that traditional methods might miss. For example, analysing URLs for slight variations that indicate a phishing attempt or identifying AI-generated scam emails." He described how phishing campaigns have become almost indistinguishable from genuine communications. "The old advice about bad grammar or strange formatting doesn't apply anymore. With AI, attackers are producing flawless emails in seconds," he said. "But the good news is that AI on the defensive side is just as powerful." Mimecast's platform uses AI throughout its stack, from sandboxing and behavioural analysis to identifying language markers in emails associated with business email compromise (BEC). "We look for those AI fingerprints - which often show up in machine-generated messages," he explained. For example, if there was an email that simulates a CEO urgently requesting staff to buy gift cards - a common BEC tactic - Mimecast's AI can intercept it. "Instead of an employee reacting to that urgency, we use AI to throw bubble wrap around them, flagging the threat before any action is taken," he said. Trust in AI is still an issue, however. "It's a double-edged sword," O'Hara acknowledged. "There's hype fatigue in cybersecurity - zero trust, now AI. And the problem is when vendors slap 'AI' onto everything, it erodes trust." He noted that some vendors rely solely on AI, which leads to high false positive rates and overburdened security teams. "AI is probability-based. Without cross-checking, it can trigger too many false alarms, and analysts burn out sifting through them," he said. "Our platform uses a layered approach - AI decisions are supported by additional checks across other systems, improving accuracy." Mimecast has gone a step further by achieving ISO certification for ethical use of AI, addressing concerns about bias and data misuse. "Transparency matters. You need to understand how the model works, especially if it goes off track," he said. "That's why we plan for machine unlearning - to rollback models if they learn something they shouldn't." Looking ahead, O'Hara envisions a future where AI acts as a sort of digital guardian angel. "Imagine a Clippy-like assistant - but useful - that knows your role, your habits, and quietly keeps you safe behind the scenes," he said. He also discussed how application programming interfaces (APIs) play a crucial role in integrating Mimecast's human risk platform with other systems. "We pull in data from HR, endpoint and identity platforms to paint a picture of risk - right down to the individual level," he explained. "If someone's on notice or switching roles, their risk profile changes. APIs help us adapt protection accordingly." Importantly, AI in cybersecurity is no longer just about detection and defence. Mimecast also uses it for prediction and prevention. "With data from 44,000 companies and billions of emails daily, our AI tools can identify emerging threats early and act before damage is done," he said. "That's where we're moving - from reactive to proactive security." But for smaller organisations, predictive security can seem out of reach. "The average Australian SMB doesn't have the budget or capacity for that level of protection," he noted. "We offer it as a service - so they benefit without the overhead." As for the future of cybersecurity training, O'Hara predicts a shift from generic instruction to highly tailored behavioural nudges. "Instead of monthly sessions, we'll see hyper-contextual, AI-generated interventions in the moment," he said. "That's the power of AI - it knows how to reach each individual in a way that resonates." He added that balancing automation with human oversight remains a key concern. "Right now, most organisations use automation to assist - not replace - analysts. And that's wise," he said. "False positives can grind a business to a halt if something like Salesforce gets blocked. But as AI improves, that balance will shift." Ultimately, he believes that the most exciting developments are still unknown. "I'm genuinely excited by what we don't yet see coming," he said. "AI has unlocked possibilities that feel like magic." And while security teams dream of AI replacing their most tedious tasks, O'Hara points out there's a long way to go. "If AI can act like Cinderella's godmother - guiding users to return home just before the stroke of midnight - then we're on the right track," he said.
Techday NZ
2 days ago
- Techday NZ
GigaChat AI assistant achieves 93% accuracy in medical diagnoses
SberHealth's GigaChat-powered artificial intelligence assistant has demonstrated a diagnostic accuracy rate of 93% during recent tests conducted by the Artificial Intelligence Research Institute (AIRI). The experiment involved the AI healthcare assistant, which is based on the GigaChat neural network model, diagnosing 30 real clinical cases that were randomly selected from the New England Journal of Medicine. These cases varied in complexity, and the testing methodology used was similar to an experiment conducted by Microsoft to verify its own AI diagnostic orchestrator, MAI-DxO. According to AIRI, the SberHealth system established correct diagnoses in 28 out of 30 cases, while a comparable foreign solution recorded an 85% accuracy rate. The AI assistant operated with limited initial data, receiving only the patient's gender, age and symptoms before interacting through simulated doctor-patient dialogues. It followed a sequence of requesting additional clinical tests, imaging, or consultation information as needed to make diagnoses. The median number of dialogue turns between the AI and the simulated patient was three, indicating a relatively high speed of decision-making. Sergey Zhdanov, Director of the Healthcare Industry Centre at Sberbank, said: "The experiment demonstrated that our technology is not only competitive but also sets new standards in medical diagnostics worldwide. We observe how multi-agent architecture speeds up and enhances the diagnostic process. It's particularly important that the system exhibits flexibility: it revises hypotheses, requests additional data, and even responds to the emotional presentation of clinical scenarios. In the future, this opens up opportunities for interdisciplinary care teams, with AI serving as a reliable assistant to physicians." During the experiment, each clinical case was labelled by level of difficulty. The AI system was able to successfully identify and diagnose several rare conditions, including Whipple disease, which it recognised in one step, aceruloplasminemia, identified in six moves, and rasburicase-induced methemoglobinemia. The assistant's performance was characterised by several features, according to researchers. It typically completed diagnoses in three moves, deployed logical reasoning, and handled both rare and complex pathologies. The system was also noted for its ability to blend clinical accuracy with a dialogue logic that could adapt effectively to different presentation styles, which included effectively responding to emotional cues in simulated scenarios. Ivan Oseledets, Chief Executive Officer of AIRI, commented: "Today, multi-agent systems are capable of confidently identifying rare, masked pathologies that go beyond typical emergency department algorithms. Can a medical AI assistant adjust its hypothesis in time, discarding the most probable but incorrect pathway? The AI assistant proved it could, doing so faster than anticipated by a seasoned observer with 15 years of medical experience." The researchers at AIRI described the experiment as exploratory and indicated that further development is planned. They have proposed expanding the sample size by incorporating additional cases from other medical journals to investigate the capabilities of the assistant more widely. The system's potential uses were not limited solely to practical medicine but also extended to the area of physician training, where it could offer realistic simulations of complex clinical cases. The GigaChat-based assistant is a product of cooperation between AIRI and SberMedAI. Since its introduction, it has been piloted in the SberHealth app and has already been used over 160,000 times in real conditions to assist people seeking medical support. Follow us on: Share on:
