It used to be easy for US Bank to determine which users and systems could be trusted, and which should be viewed with suspicion. Then along came Web 2.0.

"We always said outside the corporation was untrusted and inside the corporation was trusted territory," US Bank CTO Gary Hodge said in a keynote panel discussion on Web 2.0 security at Interop Las Vegas Wednesday. "Web 2.0 has changed all that. We've had to expose the internal workings of the corporation. There's a whole rash of new devices coming out to enable people to compute when they want to, with the iPhones and smartphones."

At the sixth-largest bank in the country, Hodge is worried. While it took a decade or more to gain a "level of hygiene" in PCs, with virus scanning and other security tools, he thinks smartphone developers haven't paid enough attention to security.

"I don't think most people have thought about their smartphone in that context," he said. "There's probably a whole rash of vulnerabilities that will show up in the next few years, and we're not sure what they're going to look like."

Hodge was joined by Gary Dobbins, director of information security at the University of Notre Dame, and two officials from the vendor Secure Computing.

Secure Computing sees more than 10,000 malware samples a day, and they are growing in sophistication as organized crime and terrorists utilize the Web for malicious purposes, said Dmitri Alperovitch, principal research scientist for Secure Computing.

"The potential from a criminal perspective has expanded dramatically in the last several years," Alperovitch said. "It's no longer about someone breaking into a computer and hacking your Web site. It's much, much more serious."

There are two main problems, according to Alperovitch. Content in the Web 2.0 world can be produced by any individual who comes to your Web page, particularly social networking sites. Secondly, the browser is now the operating system, providing access to instant messaging, Web conferencing, telephony and any other number of services. The security perimeter is shrinking, Alperovitch said.

The advice from Dobbins is simple: "Never trust the browser," he said. "It's amazing how many sharp programmers will make that mistake."

IT spent years learning how to best authenticate users to computers, Hodge of US Bank said. With the rise of malware, the problem is now authenticating software to software, or computers to computers.

In the banking world, allowing a data breach might be the surest way to wind up in court. The security perimeter expansion caused by Web 2.0 has forced US Bank to take a harder line, allowing employees to see only the information they need to do their jobs. Every device in the network is locked down, whether it's a computer, a CD or a USB drive.

"Any device that can move information out of the corporation is pretty well restricted," Hodge said. "We monitor every electronic transmission that goes out of the organization. We're expected to know, when a file leaves the organization, what its content is, who has access to it."

Too much security can harm the bottom line and annoy users, Hodge and Dobbins noted. It's a balancing act, and a dangerous one if balanced improperly.

"We protect money. It's new for us to have to protect vast amounts of information," Hodge said. "We spend millions of dollars on security but it doesn't generate any new revenue. I haven't been able to show anybody a return on investment. It comes down to can we secure the organization at the right risk and the right cost. You can't spend all the money. You have to figure out what level of risk you're willing to tolerate."