Monetizing
User Data
Context :
How India can
upend world’s data economy
HOW DO YOU PROTECT PERSONALLY IDENTIFIABLE USER
DATA? AND IF SOMEONE AGREES TO THEIR DATA BEING MONETIZED, DOES A COMPANY HAVE
TO SHARE A PART OF THE REVENUES WITH THE USER?
·
Hindustan Times ST (Mumbai) / 8
Jul 2023 / Author > Charles Assisi
·
Finally,
the Personal Data
Protection (PDP) Bill 2023 will be tabled during the monsoon session of
Parliament. This is something we were expecting to happen during the Winter
Session and had discussed earlier in October on these pages. Be that as it may,
our conversations with those who worked behind the scenes to craft the Bill
suggest that in its current avatar, this Bill holds the potential to “upend the Data Economy of the world.”
What
is that supposed to mean? Sharad Sharma, ( sharad@ispirt.in ) a volunteer at the technology think-tank iSpirt who is among those who brainstormed on
the Bill with other thought leaders in the tech ecosystem.
“If
implemented, this can create `The
India Way’ or ‘The Third Way’ on how to think about data”, he says.
The
First Way meant ignoring the problem. When it ran the course, the Second Way came into being, which is to anonymise data on a good faith basis.
But
‘good faith’ need not compel action because it may be at loggerheads with an
entity’s motive.
Consider
India alone. Almost 90% of people use the Chrome browser built by Google. It
collects data that is used by another division of Google to target users with personalized ads.
All the monies are pocketed by
the company and the user whose data is traded gets nothing.
Does
it have to be this way? At end of the day, all of this personal data is traded
for profits by a third party.
Why
shouldn’t Google be sharing a part of the profits with people whose data it
sells? And how can we be sure Google isn’t
sharing data if we don’t want it to be shared?
This
means it is time to ask the big questions: How do you protect personally identifiable user
data? And if someone agrees
to their data being monetized, does a company have a responsibility to share a
part of the revenues it earns with the user?
While
most companies argue no personally identifiable data is collected or traded,
those embedded in technology know this is untrue. In fact, back in October
2006, Reed Hastings, the co-founder of Netflix, announced the ‘Netflix Prize’.
Anyone who creates a recommendation engine which is better by 10% than their
current algorithms gets to keep $1 million. Only “anonymized data” would be
released. It wasn’t too long before coders started to create better algorithms,
but they identified individual users as well from that data. The program has
since been disbanded.
“What
it showed,” says Sharma, “is that anonymization is an inexact science.” In fact, he argues, while
a company may claim it does not collect any personally identifiable data, there
is no incentive for it not to attempt privacy violations.
And
if a user agrees to share their data, how much data must they share. Here again, Sharma explains, the
problem has been “no one knew how to restrict any
entity from collecting more than what is needed.” And this, he says, is the hairy problem Indians have cracked.
When
asked how, Sharma explains that the ‘India way’ which is coming up for scrutiny
works differently. When someone built a model, like Netflix did, people could
compare it to other datasets in the public domain.
In
the new scheme of things, after someone builds a model, it is sent to a
“Computationally Clean Room” (CCR).
The
model can only inspect patterns, and not data. This is computationally
guaranteed via a mathematical framework called Differential Privacy
(DP).
When
a DP and the CCR is merged, it becomes pretty much impossible to identify a
person. Work on this is in progress.
Why
does this matter and how does it hold the potential to upend the Data Economy
of the world?
The
first is that it opens the door to compensate people who may want to be compensated adequately for
their data.
Open Source Software will keep them
informed of how many times their data accessed and what they are owed.
The
second, Sharma says, is that between India’s large population, a legislation in
place, no ambiguity on data ownership,
the doors will begin to open up for India to begin work at becoming the “ Model
Making Capital of the world” as well.
It
appears all the ducks are lined up for India to shoot.