Would you trust this man with your online life?

Google's Peter Fleischer tells Karlin Lillingtonabout the company's privacy policies and the challenges ahead for online data protection

PETER FLEISCHER has the dubious privilege of being one of the most prominent targets in the internet privacy firing line. As global privacy counsel at Google, Fleischer is the privacy point man for a company that is both much loved (all those very cool services, plus a share price to die for) and disliked (what are they doing with all that data, and what might they do with it in future?).

In a week in which Google launched a health site where people can upload and store their personal medical records - a scenario that makes many privacy advocates blanch - Fleischer is in Dublin for an online privacy forum today at the Institute for International and European Affairs.

"This is all kind of new territory," he says. Because of Google's pivotal role as the internet's guide dog - 61.6 per cent of US internet searches last month were done on Google - and because search data from users is at the core of its business model of serving up targeted advertising, Google is in the middle of an international privacy debate.

Regularly criticised, Google has already had to shift policy and practice in the face of privacy concerns. For example, under pressure from Europe with its stricter data protection regulations, it now anonymises search data after 18-24 months.

It has been kicked for having a convoluted privacy policy, which in turn is not accessible from its search homepage. Fleischer was in the news recently for purportedly admitting the policy was "vague", but states this was only a reference to the difficulty of specifying exactly which third parties it might be required to turn personal data over to if subpoenaed.

Privacy complaints have led the company to blur faces of individuals captured in its street-level view of cities in Google Maps. Then there is the merger with advertising giant DoubleClick, which some fear will stir together massive databases to expose individuals' online private lives.

In addition, through its Gmail webmail service and ownership of blog service Blogger, Google holds e-mail and weblog data from millions of users, who can construct personal iGoogle portal sites with their diary and other information. And now, health records.

Fleischer, who has a weblog entitled Privacy . . . ?, acknowledges that Google's search service "raises all the very hard questions", which he reviews with data protection commissioners across Europe and with privacy advocates, Google users, and governments.

"The context is that we are facing a world in which people are trying to understand what privacy means on the internet," he says.

He describes as "awkward" the overall process of trying to figure out how to manage privacy issues in the dynamic, constantly evolving, but very young internet medium that has no consistent international legal framework. "Google is new to this, too," he argues.

He ticks off key challenges that must be resolved if companies - Google as well as any other internet businesses - are to flourish. All are being debated with the Article 29 group of European data protection commissioners, which published a privacy document on the topic last year.

First, the issue of jurisdiction. "Whose laws apply when you have one internet architecture?" he asks. For a US company that manages data from users all over the world, this is a central question.

"We also all agree that fundamentally, users should be in charge of their data. But how do you implement that?"

Fleischer says people also need to distinguish between individualised and personalised services.

Many online services are individualised without identifying the specific user. An example is advertisements that recognise basic geography from the IP address (location address allocated by a web user's internet service provider to their computer), but don't recognise the user as an individual connected to, say, a Gmail account and specific searches.

The Article 29 group wants to define any data associated with an IP address as falling within data protection guidelines. Google argues that, as it cannot immediately link a person with an IP address, the data shouldn't be considered personal.

He also insists that privacy considerations should not have been part of the DoubleClick merger review in the US or in Europe, because such matters are not the remit of a competition review. Privacy advocates who pushed for such a consideration "were looking for a way to leverage privacy into the debate. But that's not to deny that it's an important issue." It just should be debated in other arenas, he says.

The merger will allow for more focused advertisements, but the process is benign, he states. DoubleClick doesn't identify individuals with web visits. But maybe people fear what could be done, perhaps without them knowing or understanding the implications.

That's why international privacy protections need to be agreed, says Fleischer. "Public policy has to be developed out of global privacy standards." He points to international agreement on copyright as a model for what could be done. He also notes that data retention - the requirement that call, e-mail and internet data be retained for law enforcement - is a growing concern for the company. Google is talking to the Government here about its policies in this area on this trip, he says.

Though Fleischer is amiable, he bristles at what he sees as unfair attacks on Google's privacy record.

He says the company has very clear, explicit privacy policies. "We've also been more upfront than any company about government access to user data," he says. Of 30 companies quietly subpoenaed for user data by the US government, only Google refused to divulge its records and went to court - and won.

Does he find the criticism frustrating? "As I see it, a lot of people hold us to a higher standard than other companies. And that's fair enough. I hold us to a higher standard than other companies."