At ZDNet's IT Project Failures blog, there is a good discussion about whether or not Twitter is dangerous. I can see why corporate IT professionals probably view Twitter as dangerous, since it is another (new) way in which sensitive company data can easily be divulged by employees. To me, Twitter is no different than other computing technologies (e.g., instant messaging, e-mail, flash drives, etc.) which employees can use properly or abuse (e.g., share sensitive company data with people who shouldn't access to that data).
Companies exist by meeting the needs of their customers. If their customers use Twitter, then the company should use it, too. It is always wise to, "fish where the fish are" -- for companies to communicate with their customers based on their customers' communication preference. If their customers use Twitter, then it's the wise company that Twitters to read what their customers are saying about their brand.
Regarding data security, the bigger issue is corporate training. Several readers of the IT Project Failures post have correctly commented that since employees have signed confidentiality agreements, this should be a protection. I agree: should be. There's a big difference between signing a confidentiality agreement on day one of their employment vs. complying with agreement years later as new technologies emerge.
I see no problem with students or home users who Twitter. Just like home computer users need to learn good data security habits to protect their identity data, corporate employees need to be trained on data security threats and how to practice good data security habits at work. The large number of data breaches involving laptops is one indicator that many employees don't practice good data security habits. And, that employee training should include new technologies like Twitter for both corporate employees and corporate IT staff.
Since this blog is about identity theft and corporate responsibility, I write mostly about consumers, who are either employees or former employees affected by corporate data breaches. I can think of several good applications where Twitter is appropriate and beneficial. For example, an activist blogger can use twitter to highlight or to document their experience or a problem. Another example: a company can use Twitter as another method for customers to interact with its brands. Twitter isn't for everyone as this Twitter 101 post and Matt Dickman's video explain.
What's dangerous are companies that don't enforce effective data security policies and processes... when a company loses backup data tapes, or when databases become corrupted. What's dangerous are employees that don't enforce good data security habits.
I don't see Twitter as a problem since there are so many other ways companies lose thousands of employee and customer records during data breaches. What's dangerous are companies that suffer repeated data breaches. What's dangerous are companies that don't inform identity theft victims promptly of the data breach. What's dangerous are companies that offer free credit monitoring services to ID-theft victims, while that offer duration doesn't match the risk period created by the company's data breach.