September 30, 2025

news  
Microsoft Partners Proofpoint, Anthropic Expands Workforce

The evolving landscape of artificial intelligence continues to bring both innovation and new challenges across various sectors. In the realm of application security, Harness has acquired Qwiet AI to bolster its offerings against vulnerabilities in AI-generated code, while Legit Security has updated its AI Security Command Center to manage risks associated with AI code, models, and servers. Mplify is also addressing AI-driven cyber threats with a new SASE certification. Meanwhile, the convenience of AI travel planning tools like ChatGPT is tempered by their tendency to provide inaccurate or misleading information, prompting experts to advise users to verify AI-generated advice with reliable sources. In the automotive industry, London-based Wayve is developing an AI Driver system that mimics human driving behavior, with plans for integration into Nissan's systems starting in 2027 and trials with Uber. Tech giants are facing increased scrutiny over their substantial AI investments, with investors closely monitoring returns amidst slowing market gains. On the consumer front, a new AI dating app called Amata is launching in New York City, offering AI-planned dates for $16 and having raised $6 million to combat dating app fatigue. In cybersecurity, Proofpoint is integrating AI for threat detection and is partnering with Microsoft to enhance its security platforms. Separately, Anthropic is expanding its non-US workforce and has agreed to pay damages for copyright infringement. In Ohio, a legislative effort is underway to ban AI from being granted personhood or the ability to marry. Finally, Sage Intacct is leveraging AI to improve financial productivity for schools by automating tasks and detecting anomalies.

Key Takeaways

  • Harness acquires Qwiet AI to enhance application security against risks from AI-generated code.
  • Legit Security updates its AI Security Command Center to manage risks from AI code, models, and servers.
  • AI travel planning tools like ChatGPT can provide inaccurate information, requiring user verification.
  • Wayve AI Driver is developing AI that mimics human driving for autonomous vehicles, with integration planned for Nissan systems in 2027.
  • Tech companies face pressure to demonstrate positive returns on significant AI investments.
  • Mplify introduces a SASE certification to help enterprises address AI-driven cyber threats.
  • Amata, a new AI dating app, raises $6 million and plans AI-assisted dates for $16.
  • Proofpoint is enhancing AI in threat detection and partnering with Microsoft for integrated AI capabilities.
  • Anthropic plans to expand its non-US workforce and has agreed to pay damages for copyright theft.
  • An Ohio bill proposes to ban AI from being granted personhood or the ability to marry.

Legit Security upgrades AI Command Center for code risks

Legit Security has released an updated AI Security Command Center to help manage risks from AI-generated code. The tool provides a comprehensive view of how AI code, models, and servers are used in software development. It detects risky AI model usage and offers real-time insights into AI-related security threats. The platform also provides risk metrics for teams and applications, helping identify areas needing support.

Legit Security enhances AI Command Center against code risks

Legit Security has significantly updated its AI Security Command Center to address the risks associated with AI-generated code and AI-first development. The platform offers a detailed overview of AI code, models, and MCP server usage throughout the software development lifecycle. It identifies unauthorized or risky AI model use and provides real-time data on AI-related security dangers. The updated center also includes team and application-level risk metrics to help manage security postures.

Harness acquires Qwiet AI to boost application security

Harness has acquired Qwiet AI, a company specializing in AI-powered vulnerability detection, to strengthen its application security offerings. This move aims to address the increasing risks from AI-generated code, which often contains hidden vulnerabilities. The combined technologies will help organizations reduce alert noise, speed up remediation, and focus on critical security issues. Rahul Sood has joined Harness as General Manager to lead the application security business.

AI travel planning can lead to dangerous mistakes

Travelers are increasingly using AI tools like ChatGPT for trip planning, but these tools can sometimes provide inaccurate or even dangerous information. Examples include directing tourists to non-existent locations or providing incorrect information about transportation schedules. AI generates responses based on patterns in text, which can lead to 'hallucinations' or made-up information that is hard to distinguish from facts. Experts advise users to verify AI-generated travel advice with reliable sources.

AI travel advice often inaccurate and unreliable

While AI tools like ChatGPT offer convenience for travel planning, they frequently provide incorrect information, leading to frustrating or even dangerous situations for travelers. Studies show a high percentage of AI-generated travel advice contains errors, such as recommending closed attractions or inaccurate travel times. Experts explain that AI's text-based pattern recognition can result in confident but false recommendations. Travelers are advised to use AI as a starting point and always verify critical details with human sources.

Wayve AI Driver mimics human driving for autonomous vehicles

London-based Wayve has developed an AI Driver system that learns from millions of hours of driving data to mimic human behavior in vehicles. Demonstrated in Detroit, the AI guides a Ford Mustang Mach-E through traffic, relying on internal cameras and artificial intelligence. This technology aims to enhance advanced driver assistance systems (ADAS) and is being integrated with Nissan's systems starting in 2027. Uber is also partnering with Wayve for public-road trials of autonomous vehicles.

Tech companies face pressure on AI investment returns

Big Tech companies are under pressure to ensure their significant AI investments yield positive results, as market gains driven by AI trends show signs of slowing. While cloud giants are seeing increased revenue from AI workflows, high capital expenditures carry risks. Strategists suggest rebalancing portfolios into sectors like healthcare and small caps, which may offer better value. Investors are closely watching if these massive AI spending plans will translate into sustainable revenue and profits.

Mplify launches SASE certification for AI cyberthreats

Mplify has introduced a SASE (Secure Access Service Edge) certification to help enterprises address AI-driven cyber threats. The new SASE certification validates solutions against tested standards for network security, security service edge, and SD-WAN. This initiative, backed by major companies, aims to provide clarity and confidence in implementing secure services across enterprise domains. Mplify's Enterprise Leadership Council developed the manifesto to combat fragmentation in cybersecurity planning and promote secure AI operations.

AI security news from the week of September 22, 2025

This week's security and AI news includes data breaches affecting Jaguar Land Rover and Harrods, and a proposed UK Digital ID card raising privacy concerns. National agencies warned of a sophisticated espionage campaign targeting Cisco software. Denmark and other countries experienced drone disruptions, possibly linked to Russia. Anthropic plans to expand its non-US workforce and agreed to pay damages for copyright theft. Cloudflare Inc. announced plans to hire interns and offer free development tools to universities.

Sage Intacct uses AI to boost school finance productivity

Sage Intacct, a cloud-based accounting software, is using AI to improve financial processes for schools. The software automates tasks like invoice processing and coding, reducing manual data entry and errors. Its GL Outlier Detection feature uses machine learning to flag unusual journal entries, helping to safeguard financial integrity. These AI-powered tools aim to provide real-time data, improve accuracy, and allow finance teams to focus more on analysis rather than manual tasks.

New AI dating app Amata plans dates and skips swiping

Amata is a new AI-powered dating app launching in New York City that aims to connect users in real life quickly. Instead of swiping, users chat with an AI matchmaker that suggests potential dates. For $16, the AI also plans the first date, including location selection. The app limits direct messaging to two hours before the date to encourage in-person meetings. Amata has raised $6 million and seeks to address dating app burnout by focusing on intentionality.

Proofpoint secures AI interactions with new tools

Proofpoint is enhancing enterprise security by focusing on AI in threat detection and human interactions. At Proofpoint Protect 2025, the company highlighted its use of AI models to analyze emails, treating each as a potential prompt for phishing attacks. Proofpoint Satori Agents are introduced to support Security Operations Center (SOC) teams by automating tasks like data loss prevention and phishing simulations. The company is also forming partnerships with Microsoft and CrowdStrike to integrate AI capabilities across security platforms.

Ohio bill bans AI marriage and personhood

An Ohio lawmaker has introduced House Bill 469, which would prevent artificial intelligence systems from being granted personhood and prohibit people from marrying AI. The bill declares AI systems as non-sentient entities, banning them from holding corporate roles or owning assets. Representative Thad Claggett stated the bill aims to maintain a clear separation between humans and machines. The legislation addresses growing concerns about AI's rapid evolution and its potential impact on society.

Sources

AI Security AI-generated code Vulnerability detection Application security AI travel planning AI hallucinations Autonomous vehicles AI Driver system AI investment Cybersecurity SASE certification AI cyberthreats Data breaches Privacy concerns Copyright infringement School finance AI productivity Dating apps AI matchmaker Threat detection Phishing attacks AI personhood AI ethics