AI agents are beginning to surpass human employees in many high-tech office scenarios, and they are also more difficult to monitor.
AI Security and Governance Startup Opsin announced on Thursday that it has raised $21 million to address this trend. As corporate employees integrate more and more AI from companies such as OpenAI and Microsoft into their work processes, this company is helping businesses manage these new "intruders" that may pose risks.
For non-technical employees, deploying agents – a type of software based on AI that can complete tasks autonomously – has never been easier than it is now. As employees integrate this kind of software into the company's work processes, code, and databases, they may gradually lose control over what AI is actually doing.
Opsin, the Chief Executive Officer and Co-Founder, told Business Insider that when he realized security concerns were slowing down the development of AI, he decided it was time to expand the company.
“agents is actually more human than you might imagine.”Pham said,The reason is that you don't know what's going on in a person's mind, but you can judge by their behavior.
The goal of Opsin is to provide a comprehensive view to help enterprises understand what AI and agents are actually doing within the organization, as well as the potential risks that these actions may bring. SYN Ventures and Race Capital led this round of financing. The company currently has about 20 employees, and this funding will be used to expand the sales and engineering teams.
Security professionals need to clean up the chaos caused by AI and agents.
Pham is a graduate of MIT. He and co-founder Oz Wasserman worked together at Abnormal Security, and both left the company in 2022. When deciding to start their own business, Pham mentioned that they interviewed hundreds of people to understand what companies were most concerned about when introducing AI.
The issue of sensitive data is the most prominent. Pham said that almost everyone mentioned that AI and agents might access and abuse personal or confidential information.
Today, Opsin helps customers understand where agents is active, which tools and data they have accessed, what they were originally designed for, and which agents may need to be cleaned up. Subsequently, the company will also continue to monitor the corporate work environment, observe agents, and flag risky behaviors.
“In the end, the key lies in the behavior of the users, the behavior of agent,”Pham says.
He said that even in companies with relatively weaker technical capabilities, the scale of agents still shocks some customers. Pham indicates that in a company that uses Opsin services, a chief financial officer created a agent that inadvertently granted other employees the highest level of access to financial information.
The cyber threats associated with AI may bring about prosperity.
The cybersecurity industry has always been in a state of high activity, and the growth of agents has added fuel to this trend. SYN Ventures, a cybersecurity investor that supports Opsin, told Business Insider that as more enterprises adopt AI products, hackers will also find more targets.
The more people do this, the easier it is for attackers to gain access and carry out attacks.Chisholm says.
Network defense stocks have seen a significant increase this year, and Chisholm says he has noticed an increase in venture capital as well. Both Anthropic and OpenAI are actively recruiting cybersecurity talent.
Opsin aims to make enterprises feel more secure when using Microsoft Copilot, ChatGPT Enterprise, and other products. Chisholm holds an optimistic attitude towards the industry's adaptation to AI.
"In the coming years, what we need to do is learn how to make the most of it, rather than simply fearing the concept itself."He said.
