Google Patches AI-Discovered Chrome Vulns; Pluto Raises $42M

Key Takeaways

• Lightspeed AI launched agentic retail tools on October 15, 2026, with all actions requiring retailer approval and logged for audit.
• Pluto Security raised $42 million to secure enterprise AI tools, with finance teams identified as the largest source of risk.
• Google paid XBOW $250,000 after AI-assisted research found three Chrome vulnerabilities, which were patched in August and September updates.
• Hackers used DALL-E to create harmful content, and Stability AI said it is not responsible for outputs generated by the model.
• Black Sesame Technologies signed a RMB 300 million AI server deal, its first for the AI computing server business.
• Pluto Security has over 30 enterprise customers and more than $2 million in annual recurring revenue.
• Lightspeed AI's agent features are in opt-in beta with broader availability coming soon.
• Black Sesame's server deal involves about RMB 250 million in complete systems using its own high-performance AI chips.
• Google confirmed and fixed the Chrome bugs found through AI-assisted testing, and the exploit was never used in the wild.
• Thailand's privacy regulator released non-binding AI guidelines covering training data and model deletion obligations.

Lightspeed AI launches agentic tools for retailers

Lightspeed AI introduced new agentic capabilities for retail on October 15, 2026. Retailers can ask questions, get recommendations, and prepare tasks like tagging clearance items or planning inventory transfers. Every action requires retailer approval and is recorded in an audit log. The features are in an opt-in beta with broader availability coming soon. Dax Dasilva and Bhawna Singh of Lightspeed explained the goal is to simplify routine work while keeping retailers in control.

Lightspeed AI brings retail insights into one workflow

Lightspeed AI now combines catalog, inventory, sales and purchasing data across retail locations. The agent can identify low-stock products, prepare purchase orders and move inventory to where demand is strongest. Retailers see a preview of each action before it is taken. Dax Dasilva, Founder and CEO, said the new features remove routine work from retailers' plates. Bhawna Singh, CTO, said the agent works within systems merchants already use every day.

Lightspeed AI adds agent features in opt-in beta

Lightspeed AI rolled out new agentic capabilities on October 15, 2026 for select retailers. The agent uses operational context from physical retail including inventory and sales data. It can update product data, surface sales opportunities and prepare tasks for approval. An audit log tracks every action taken. The new features are being introduced through an opt-in beta with broader availability coming soon.

Black Sesame Technologies signs RMB 300 million server deal

Black Sesame Technologies signed a sales framework agreement for AI computing servers worth about RMB 300 million on October 14, 2026. An A-share-listed company in China will purchase around RMB 250 million of complete AI computing server systems built on Black Sesame's high-performance AI chips within one year. This is the first order for Black Sesame's AI computing server business. CEO Shan Jizhang said the deal extends the company's AI computing platform into the server market.

Black Sesame Technologies lands first AI server order

Black Sesame Technologies announced a RMB 300 million AI computing server sales contract on October 14, 2026. The partner will buy roughly RMB 250 million of server systems using Black Sesame's high-performance AI chips within one year. The deal marks the first order for Black Sesame's AI computing server business. AI computing demand is growing beyond cloud data centers into edge and on-premises applications. The company plans to pursue further AI computing service partnerships as demand expands.

Thailand issues AI privacy guidelines for training data and model deletion

Thailand's privacy regulator released new non-binding guidance on AI and personal data. The guidelines say AI models trained on personal data may still fall under the Personal Data Protection Act if individual information can be extracted. They also cover using publicly available data for AI training and handling deletion requests for trained models. The regulator stressed that technical difficulty in removing data from AI models does not remove legal obligations.

Local cybersecurity experts use AI to counter AI driven cybercrime

Cybersecurity experts in Toledo, Ohio say criminals are using AI to find vulnerabilities and create convincing attacks against everyday people and businesses. Experts Gwen Betts and Jason Slagle urge people to stay suspicious and think critically before acting on requests for money or sensitive information. They warn that AI makes cybercrime expertise available to more people and that automated responses to suspicious emails can lead to fraud. The experts say defending against these threats remains a constant cat and mouse game with technology.

Human Advantage Conference 2026 explores Behavioural Science and AI

Ipsos announced the Human Advantage Conference 2026, which will bring together leaders in Behavioural Science, AI and Human Intelligence. The event will feature speakers, workshops, panel discussions and networking opportunities. Organizers say the conference will focus on using these fields to drive business growth and improve decision making. Details on the specific date, location and speakers are listed as placeholders in the announcement.

Pluto Security raises $42M to protect enterprise AI tools

Pluto Security came out of stealth with $42 million in funding to help companies secure the AI tools their employees use. The funding includes an $8.5 million seed round and a $34 million Series A led by Jump Capital. CEO Shahar Bahat said about 60% of detected risks come from tools built by non-developers, with finance teams being the largest source. The company has over 30 enterprise customers including Oscar Health and Sixt and has reached more than $2 million in annual recurring revenue.

Google pays $250,000 after AI helped break Chrome security

Google paid security firm XBOW $250,000 after researchers used OpenAI's GPT-5.6 Sol to help find Chrome vulnerabilities. The team used AI to search Chrome's source code, test weaknesses and build exploit code while humans chose the targets and reviewed the results. The exploit combined three Chrome flaws involving V8, WebP and Crashpad to escape the browser sandbox on Windows. Google confirmed the bugs and fixed them in the August and September Chrome updates, and XBOW said the exploit was never used against real users.

MLS leaders debate AI data access and governance

Real estate MLS executives disagree on how AI should use listing data. eXp Realty wants Model Context Protocols for real time data access. Compass CEO Robert Reffkin prefers keeping listings protected. Experts like Nick Aufenkamp use AI tools to analyze large sets of MLS data. Northstar MLS and CRMLS are working on monitoring tools with WAV Group's Fluente AI.

AlphaRLM uses reinforcement learning to discover financial factors

AlphaRLM is a reinforcement learning framework that trains language models to build symbolic factors from financial literature. It uses LEGO to guide factor construction with literature based ideas and market feedback. Tests on CSI 300, CSI 1000, and S&P 500 show strong results. The 8B model performed well across different markets.

AI companies face criticism over harmful AI generated content

Hackers used the DALL-E AI model to create disturbing and violent content. Dr. Timnit Gebru reported the issue and raised concerns about AI safety. Stability AI said it is not responsible for content generated by the model. Experts argue that AI companies need to take more responsibility for the messes their tools create.

Sources

• Lightspeed AI Expands From Insights to Action With New Agentic Capabilities for Retail
• Lightspeed AI Expands From Insights to Action With New Agentic Capabilities for Retail
• Lightspeed AI Expands From Insights to Action With New Agentic Capabilities for Retail
• Black Sesame Technologies Signs RMB 300 Million AI Computing Server Sales Contract
• Black Sesame Technologies Signs RMB 300 Million AI Computing Server Sales Contract
• Thai AI guidelines address privacy risks in training data, model deletion | MLex | Specialist news and analysis on legal risk and regulation
• Local cybersecurity experts fight AI threats with AI counter measures
• The Human Advantage Conference 2026: Navigating a complex world through Behavioural Science, AI and Human Intelligence
• Pluto Security Emerges From Stealth With $42M to Secure Enterprise AI Agents
• Google paid $250,000 after researchers used OpenAI’s AI to break Chrome’s security
• MLSs race to set AI guardrails before data slips beyond their control
• AlphaRLM: Aligning Language Models for Literature-Grounded Symbolic Factor Discovery
• AI Hacks Leave a Mess That AI Companies Aren’t Paying For