Microsoft’s TikTok deal is a “poisoned chalice”, says Bill Gates

Tiktok logo in front of US flag.
Donald Trump’s trade war with China has forced TikTok’s parent to sell the site. Photograph: Herwin Bahar/ZUMA Wire/REX/Shutterstock through The Guardian.

What happened

So what

  • If the deal goes through, it’ll be Microsoft’s first foray into social media and it’ll have to contend with the complexities of content moderation on a social platform.
  • Bill Gates: Facebook having some more competition is “probably a good thing” but that “having Trump kill off the only competitor, it’s pretty bizarre.”

What next

  • To avoid the backlash faced by both Twitter and Facebook, Microsoft needs to adopt a robust content moderation policy that actively protects the most vulnerable members of the TikTok community while supporting the voices of those who have historically gone unheard. So far, the bulk of conversation centers on how you respond to bad behaviour but at IF we are equally interested in designing technology to uplift marginalised groups.

Uber and Lyft must classify drivers as employees, California judge rules

Protester holding a sign with Uber and Lyft's logos that reads: we are fighting for our rights, together we are stronger.
Photograph: Getty Images through BBC News.

What happened

So what

  • Gig economy drivers in California may gain much needed economic protection during an economic depression.
  • Similar laws can be passed across the US and the world.
  • Uber and Lyft aren’t tech companies or information society services, they’re essentially taxi companies. They can’t be regarded as “intermediaries” between drivers and passengers and try to undercut the taxi industry by avoiding taking responsibility for the people who use their apps.

What next

Faculty logo on a black background

What happened

So what

  • Mass collection and analysis of citizens’ public tweets is a grey area legally and close to political surveillance.
  • The whole process lacked transparency both in how bids were assessed and in terms of the AI employed for analysis.
  • The UK population that is active on Twitter isn’t representative of the UK population.
  • The success of the privacy preserving techniques employed are unclear.

What Next

  • Though Faculty AI claims that it was ‘impossible to use the information to profile any individual or group of people’, we know that stripping out identifiers does not mean people cannot be reidentified in the future. We want to understand exactly how they protected the data and explore opportunities for applying other privacy-preserving techniques.
  • We would love to work with the relevant government teams to apply these techniques so that data can be used for public value without risking people’s privacy. We’ve already started work on this with the collection of mobility data.

Thanks to Georgina Bourke for her help editing this post.