Product Promotion Network


The scientist who gave Cambridge Analytica its Facebook data got lousy reviews online

Aleksandr Kogan, a Cambridge University academic, has long looked fishy to Amazon Mechanical Turk workers. They reviewed him negatively on a site called Turkopticon, complaining that his requests violated the service’s policy by asking them to provide access to their Facebook profiles. Bad reviews: Numerous reviews from 2014 complain that Kogan sent out work requests–which are described alternately as having people use an app or fill out a survey–asking people to offer up Facebook likes, friend lists, and some private messages.

One review says the request stated that any collected data would remain anonymous and would be used only for research. Mechanical Turk policy prohibits those using the service from making requests that collect “personally identifiable information.” No thanks: Rochelle LaPlante, a Mechanical Turk worker and advocate for ethical digital labor practices, turned down one of these requests in 2014–it asked her to fill out a personality survey–after realizing it wanted her Facebook profile information.

“It’s been in my mind since 2014 that something’s not right here,” she says. Why it matters: Kogan accessed tons of Facebook user data by creating a personality-quiz app in 2013 for research purposes; he’s an important piece of Facebook’s current scandal concerning how this data was then passed on to data broker Cambridge Analytica. What’s happening now: On Wednesday afternoon, Mark Zuckerberg broke his days-long silence about the debacle, saying Facebook is “working with regulators as they investigate what happened.”

Meanwhile, Cambridge Analytica has agreed to a forensic audit conducted by a company Facebook hired, and Cambridge University is looking into Kogan’s actions.

Image credit:

  • Turkopticon

Europe proposes taxing tech giants to ensure they pay ‘their fair share’

The European Commission has proposed a plan to add a 3 percent tax on the revenues of digital tech companies like Google, Facebook, and Amazon in order to make sure that they pay “their fair share.”

The proposal is actually a two-part plan: the revenue tax is intended as an initial measure, with plans for more long-term reforms that would eventually tax digital companies based on where they make sales, not where their physical locations are.

The proposal notes that digital companies have become far more important in recent years: nine of the world’s top 20 companies by market cap are now digitally based, versus just one in 20 a decade ago. The proposal also claims that digital-based companies have an average effective tax rate roughly half of what more traditional companies pay since EU member states can’t tax digital companies’ sales in Europe when they have no physical presence there. It’s worth pointing out that this plan would only target larger companies; smaller startups with lower revenues wouldn’t be included.

The odds of the plan actually passing seem somewhat low.

All EU member states would have to approve the proposal before it could become law, and it’s not entirely clear if countries like Luxembourg or Ireland, which benefit by serving as tax havens for US companies, would want to see major tax reform that makes it harder for companies to hide their sales from taxation.

But if nothing else, the proposal is indicative of the larger trend that we’ve been seeing in the tech world recently — that major technology companies need to be held more accountable.

Even if this proposal doesn’t pass, it at least starts to put pressure on these companies going forward.

IBM will help any firm build its own voice assistant

The firm’s Watson AI is being wheeled out (again) to imbue any firm’s device with abilities like those of Siri or Alexa–but it’s unclear if that’s really a good idea for users. The news: Engadget reports that IBM is launching a kind of white-label software service that will let other companies build voice control into anything from a smart speaker in a hotel room to the dashboard of a Maserati. The assistant will allow users to chat via typed text in some situations, too.

Why it’s interesting: The offer, really, is to allow firms to break free from making apps that run on voice software built by Google and Amazon. Companies will like that they can use their own training data, inject their own brand experience, and keep utterances away from the ears of another firm. But: It’s not clear that a surfeit of smart assistants is such a good idea for users.

Dealing with multiple AI bots on a day-to-day basis could become annoying.

And if every firm is building its own assistants without the resources of, say, Amazon, they could end up being rather less good than Alexa.

Image credit:

  • Harman / Erik John Ricardo

1 2 3 153