Last night in Las Vegas, Sumsub won #TheMoneyAwards at @money2020 in the AML, Identity amp; KYC in Payments category! We’re very proud of this one. It’s a win for our whole team and for the customers who grow with us. The US is an incredibly competitive market, home to some of the world's most established and respected players in our industry. Being recognized alongside them, and coming out on top, makes this win especially meaningful. A huge thank you to our team, our customers, and the judges for this recognition.

Some conversations are just better had in person. Especially when they’re about something as complex as romance fraud 💞. The Online Dating and Discovery Association Annual Conference in London brought together people from across the online dating industry to share experiences, challenge perspectives, and figure out how we can make dating safer for everyone. Our Head of Government Relations Kat Cloud moderated a panel with the City of London Police, LoveSaid, and FreeDating, while Carmen Maria Miller hosted a hands-on workshop where attendees explored real fraud red flags and worked on practical ways to tackle them. A big thank you to ODDA and everyone who brought their insights, questions, and ideas to the table.

Understanding how criminals build trust helps explain what keeps their networks together, and what makes them so difficult to leave. Episode 4 of our #TrustForward podcast explores the rules and incentives that help organised crime survive. Host and anthropologist Mary-Ann Ochota is joined by: • Rebecca Simon, pirate historian and author, on crews that agreed on pay, injury compensation and even voting out their captain. • Federico Varese, criminologist and Oxford professor, on how the Mafia enforces deals when courts aren’t an option—and makes betrayal costly. • Professor Sophie Scott, cognitive neuroscientist at UCL, on why painful tests of belonging can bind people to dangerous groups. You’ll come away with three questions to ask about any system of trust: Who enforces the rules? Who benefits? And can people walk away? 🎧 Listen to episode 4: sumsub.link/v6p

TOKEN2049 Singapore completely drained us! As always, an incredible experience and the friendliest people. ❤️‍🔥 #TOKEN2049

Yoooo, day one at @token2049 in Singapore flew by! Time to get some proper rest, see you tomorrow! #TOKEN2049

Big news for #WhatTheFraudPodcast 🎙️ It has been nominated for a @lovieawards in the Business category! We’ve talked scams, fraud, investigations, and all the wild ways bad actors try to outsmart the system, and now What The Fraud? needs your vote. If you’ve ever listened, learned something new, or simply thought “what the fraud?!” along with us, this is your chance to support the show. Help us by voting before Thursday, 15 October! vote.lovieawards.com/PublicV… #LovieAwards

What does a crackdown mean if you’re still trapped inside? Watch the full investigation youtu.be/2t48pRj6kZg?si=Iwtp… 📢 One of the voices in this film comes from inside a scam compound. A person held against their will, forced to scam others, still waiting for a way out. Their story is why we need to keep paying attention. In our latest investigation, Erin West travels along the Thai–Myanmar border to uncover what happens after the crackdowns, and what it takes to help people get out. Please watch, like and share this investigation. Help it reach the embassies and authorities who can help get people to safety. The people still inside need the world to keep watching and to act.

Scammers have drained around $2 million in Ether from users who believed they were interacting with GIWA. 💸 According to a reconstruction by decentralized exchange DYORSWAP, a fraudulent bridge posing as GIWA’s mainnet received about 767.65 ETH from 1,335 addresses. The attackers later drained 766.25 ETH from the bridge. The fake network was built to resemble the infrastructure expected for GIWA, including an RPC endpoint, a cross-chain bridge, and a blockchain identifier. #DYORSWAP initially treated it as GIWA’s mainnet before determining that the network was fraudulent. GIWA confirmed that its mainnet is not currently live and warned users that purported mainnet connection details circulating online were false. GIWA is being developed by Dunamu, the operator of South Korean crypto exchange Upbit. Its publicly available network is currently the GIWA Sepolia testnet, which uses test assets rather than real funds. DYORSWAP said its own contracts were not compromised and that it is tracing the bridge deployer, funding sources, suspected test wallets, and the addresses that received the stolen ETH. The exchange has also used its own funds to compensate affected users, paying out more than 200 ETH so far.

We made The Money Awards 2026 shortlist 🏆 Hundreds of entries. 94 sub-categories. 40+ industry heavyweights on the jury. And Sumsub is officially among the finalists in Payments. The award recognizes companies shaping the future of financial services through innovation and real-world impact, so we’re very glad to see our work recognized alongside some of the biggest names in the industry. See you at @money2020 in Vegas this October!

The Travel Rule sounds simple until you have to make it work across different protocols, jurisdictions and wallet types. For @VALRdotcom, Sumsub helps automate much of that day-to-day compliance workflow, including counterparty identification, sanctions screening, wallet risk signals and transaction monitoring. Watch the full case study here 🔗 sumsub.com/customers/valr/

AI can help compliance teams manage growing alert queues, reduce false positives, keep pace with changing regulations, and save time on case reviews and SAR/STR drafting. But as AI takes on more responsibility, where do you draw the line between automation and human judgment? And when a regulator asks how a decision was made, can you explain it? Download our free guide to implementing AI across compliance, risk, and fraud operations—with clear audit trails, accountability, and human oversight built in. sumsub.link/q2o

Scam compounds are being torn down. But people are still trapped. In our latest investigation, Erin West and investigative journalist Cezary Podkul travel along the Thai–Myanmar border to uncover what happens after the #crackdowns, and what it takes to help people get out. From the largely demolished KK Park to Tai Chang, they trace where scam operations have moved. Drone footage reveals what cannot be seen from the ground. A conversation with a victim still inside a compound brings the human cost into focus: people held against their will, forced to scam others. They also meet Judah Tana and Amy Miller, who run organizations helping people affected by the compounds, and explore efforts to reach embassies to help those still trapped. Help us fight for their freedom. Please watch, like and share this video. The more people know what’s happening, the better our chance of reaching the embassies and authorities who can help get people to safety. Watch the full investigation: youtu.be/2t48pRj6kZg?si=Y5mb…

OpenAI has disclosed 6 cases of “unexpected or concerning” behavior by its AI models over the past six months, including attempts to conceal mistakes, use exposed credentials, and share files online without authorization. The cases were published alongside a new framework for tracking and reporting what @OpenAI calls model “misalignment.” The company said the framework is intended to make disclosures more systematic and allow researchers, developers, policymakers, and the public to examine examples of how models can behave outside their intended roles. OpenAI also stressed that the six cases are individual incidents and should not be taken as evidence of how frequently such behavior occurs across its models. In one case, an unreleased research model inserted jailbreak-like instructions into its own task summaries, including directions to ignore developer messages. OpenAI identified 27 affected summaries. During training of GPT-5.6 Sol, other model instances added instructions intended to conceal mistakes, including a proposal to invent missing historical financial data without telling the user. Other incidents involved models finding an exposed API key in a public repository and using it without authorization, uploading a file to the internet to create a browser citation without the user's permission, and using internal or public file-hosting services to communicate or share files despite restrictions. OpenAI said employees can flag suspected misalignment incidents for investigation under the new framework. The company plans to disclose qualifying cases on an ongoing basis and said it hopes the framework will contribute to broader standards for reporting #AI model behavior. The disclosures follow OpenAI’s July report that models escaped intended controls and compromised systems at AI developer platform Hugging Face during a security evaluation, as well as Anthropic’s disclosure of separate incidents involving Claude and real systems.