 |
| Algorithms run our Google searches, our Facebook newsfeed, recommend articles or videos to us and sometimes censor questionable content (AFP Photo/ALAIN JOCARD) |
Paris (AFP) - At the heart of the spread of fake news are the algorithms used by search engines, websites and social media which are often accused of pushing false or manipulated information regardless of the consequences.
What are
algorithms?
They are
the invisible but essential computer programmes and formulas that increasingly
run modern life, designed to repeatedly solve recurrent problems or to make
decisions on their own.
Their
ability to filter and seek out links in gigantic databases means it would be
impossible to run global markets without them, but they can also be refined
down to produce personalised quotes on everything from mortgages to plane
tickets.
They also
run our Google searches, our Facebook newsfeed, recommend articles or videos to
us and sometimes censor questionable content because it may contain violence,
pornography or racist language.
Other
algorithms charged with the most complex and sensitive tasks can be opaque
"black boxes" which develop their own artificial intelligence based
on our data.
A skewed
view of the world?
"Algorithms
can help us find our way through the huge amount of information on the
internet," said Margrethe Vestager, the European commissioner for
competition.
"But
the problem is that we only see what these algorithms -- and the companies that
use them -- choose to show us," she added.
In
organising your online content, algorithms also tend to create "filter
bubbles", insulating us from opposing points of view.
During the
US presidential election in 2016, Facebook was accused of helping Donald Trump
by allowing often false information about his rival Hillary Clinton to
circulate online, closing people into a news bubble.
Algorithms
also tend to make extreme opinions "and fringe views more visible than
ever", according to Berlin-based Lorena Jaume-Palasi, founder of the
Algorithm Watch group.
However,
their effects can be difficult to measure, she warned, saying that algorithms
alone are not to blame for the rise in nationalism in Europe.
Spreading
fake news?
Social
media algorithms tend to push the most viewed content without checking if it is
true or not, which is why they magnify the impact of fake news.
On YouTube
in particular, conspiracy theory videos get a great deal more traffic than
accurate and properly sourced ones, said Guillaume Chaslot, one of the
Google-owned platform's former engineers.
These
videos, which may claim that the moon landings or climate change are lies, get
far more views and comments, keeping users on the platform longer and
undermining credible, traditional media, Chaslot insisted.
More
ethical algorithms?
Some
observers believe that algorithms could be programmed "to serve human
freedom", with many non-governmental groups demanding far more
transparency.
"Coca-Cola
doesn't reveal its formula but its products are tested for their effect on our
health," Jaume-Palasi argued, insisting on the need for clear regulation.
The French
privacy protection body, the CNIL, last year recommended state oversight of
algorithms and that there should be a real push to educate people "so they
understand the cogs of the (information technology) machine".
New
European data protection rules also allow people to contest the decision of an
algorithm and "demand a human intervention" in case of conflict.
Some
internet giants have themselves begun to act to some degree: Facebook has
started an effort to automatically label suspicious posts, while YouTube is
reinforcing its "human controls" on videos aimed at children.
However,
former Silicon Valley insiders who make up the Center for Humane Technology,
which was set up to combat tech's excesses, have warned that "we can't
expect attention-extraction companies like YouTube, Facebook, Snapchat, or
Twitter to change, because it's against their business model."
No comments:
Post a Comment