In the Age of AI, Everything Comes Down to Trust | Special Interview

Artificial intelligence has radically lowered the barrier to building technology — but it has also made it easier than ever to impersonate people, manipulate information and expose sensitive intellectual property. For Michael Ben-Shimon, founder of M&B IP, an intellectual property law firm specializing in technology companies, that makes one thing increasingly important: trust. 

“Everything is about trust,” Ben-Shimon says. “How do I trust the data and the information – and when it comes to identity, how do I know that someone trying to get into my system is really who they claim to be?"

M&B IP works with technology companies on patents, trademarks, trade secrets and copyrights, helping startups protect the technology and ideas that underpin their businesses. Ben-Shimon also invests as an angel in cybersecurity and AI companies, giving him a perspective from both sides of the equation.

“AI amplifies and accelerates both the problem and the solution,” he says. “Generative and agentic AI give attackers powerful new tools to create fake content and identities. At the same time, many smart founders and entrepreneurs are trying to solve these problems and protect enterprises against such attacks.”

AI is changing the IP equation

The explosion of AI startups has created another challenge: differentiation. Several years ago, Ben-Shimon says, there might have been only a handful of companies trying to solve a particular cybersecurity problem. “Now, we're looking at hundreds of companies in the same domain.”

Unsurprisingly, he sees intellectual property as an important moat. “Companies with strong IP can show buyers or strategic partners that what they built is sophisticated enough, and that they have a moat they can protect using IP.”

But AI is also changing the meaning of authorship and ownership. What happens when the technology itself generates the idea, code or content?

Copyright, for example, can become complicated when AI generates an image or video: “I cannot claim my own copyright because it's generated by AI,” Ben-Shimon explains, while also warning that the content could infringe someone else’s copyright.

Software development is different. Using AI coding tools such as Claude Code or Codex, he says, does not necessarily undermine the underlying IP.

“If I use a vibe coding tool like Claude Code or Codex, it's just a tool,” he says. “In IP, what we protect is the idea, the conception of the idea, and not how the idea is implemented.”

But protecting IP is only part of the challenge. Technology companies are also increasingly putting sensitive information into AI systems themselves.

“Yes, 100%,” Ben-Shimon says when asked whether companies could inadvertently give away their intellectual property by using AI. “I could ask a question and receive an answer based on confidential information belonging to someone else.”

His advice to startups is to establish an internal AI policy governing which tools employees can use, what information they can provide and how prompts are handled.

Who is really behind the agent?

As AI agents increasingly act on our behalf, identity itself becomes harder to establish.

Granting an AI agent access to personal email, enterprise systems, or confidential databases risks opening an unprecedented attack surface. “An agent is kind of live code that someone can hack and use to get information, and then use that to penetrate other systems using my same credentials,” Ben-Shimon explains.

Ben-Shimon points to Israeli startup NewCore, an M&B IP client, which is working on securing identities and access in an era of AI agents. He also highlights another client, Frame Security, which takes a different approach: training employees to recognize increasingly convincing fake content.

The threat is already unsettlingly realistic.

“They can troll you. For example, your boss can call you and tell you to cancel this entire event you’ve been working so hard on – and you’ll never realize this wasn’t actually your boss at all, but a malicious actor.”

Asked about the growing challenge of identity and trust in cybersecurity, Ben-Shimon compares the attacker-defender race to war.

“The tools, the weapons, are easy to use. To run an attack, you just need to give a few agents instructions, and that's it.”

The startup factory effect

Yet, Ben-Shimon sees a major opportunity — particularly for Israel.

“Since there are so many threats evolving so quickly, I think there will be a web of opportunities in cybersecurity, and Israel has the best of the best in cyber, deep tech and defense tech,” he says. He notes that the experience gained during the past three years of war has given Israeli companies something difficult to replicate elsewhere: technology tested in real-world conditions.

“I believe that more money will come to Israel, as investors increasingly recognize the opportunity.” 

And there is another, less obvious effect. Successful exits don't simply create wealthy founders - they create the conditions for new founders to emerge. 

“Wiz, for example, another client of ours, had a $32 billion exit,” Ben-Shimon says. “Now there’s a whole group of engineers and product managers who have the financial cushion and the opportunity to pursue their own dream of starting their own business. They don’t need to worry about their paycheck and they have the prestige of working for a very successful company.

“The more companies have successful exits, the more new opportunities are created. That's another benefit of exits that no one talks about.”

For Ben-Shimon, that may ultimately be one of Israel’s greatest advantages: an ecosystem that continually produces its next generation of companies — many of them built to solve the problems AI is creating today.

Michael Ben-Shimon will speak at CISOs. Founders. Investors., an exclusive tech forum with AWS exploring fake, trust & identity in the age of AI, on September 16 in NYC. For information and registration

follow us -Google News