Friday, September 27, 2024

LinkedIn’s feeble opt-out

Hi everyone, it's Ellen in San Francisco. LinkedIn's AI-training kerfuffle is a stark reminder that telling users they can "opt out" of some
by Ellen Huet

LinkedIn's AI-training kerfuffle is a stark reminder that telling users they can "opt out" of something is mostly meaningless. But first...

Three things you need to know today:

• Dish and DirecTV are close to agreeing on a merger
• Staff morale is plunging at SAP as the German software company restructures its workforce
• Super Micro shares fell after reports that prosecutors are probing the server maker's accounting practices

The power of default settings

Last week, LinkedIn users noticed they had been quietly opted into a new setting. Their data was being used to train LinkedIn's generative artificial intelligence models.

Users were upset that the Microsoft Corp.-owned professional network was taking their information by default without asking, and they began posting about how to opt out.

Sure, you can opt out, which sounds nice, but the vast majority of users won't end up doing it. It's that requirement to opt out that is the real issue.

Default settings have incredible power. Behavioral scientists have found that if it takes effort to change a setting — like logging into LinkedIn, or searching for the right box to un-click — then fewer people will do it.

For example, users can bolster their online security through multifactor authentication, but the default is usually password-only. In 2021, Twitter released a report that said only 2.3% of users had gone through the process to turn on the stronger protections.

"If something isn't part of the default settings, it simply doesn't exist for most people," said Rachel Tobac, the chief executive officer of SocialProof Security. "Most people don't know the choice exists."

Default settings often favor the company, not the user. And LinkedIn isn't alone: Elon Musk's social network X also auto-opted users into sharing data to train its AI, and Meta Platforms Inc. recently said it has been taking Australian user data for training. These tech companies benefit from having more training data, not less.

LinkedIn spokesman Greg Snapper told the Washington Post that the company uses people's data to train AI to "help people all over the world create economic opportunity" by improving tools to help them find new jobs and learn new skills. "If we get this right, we can help a lot of people at scale," he said.

Lax privacy defaults can also be embarrassing. Just ask Republican vice presidential candidate JD Vance. In July, Wired reporters found his Venmo account, which was publicly broadcasting sensitive information like his friends list.

Venmo's settings make your transactions social and public by default. It's particularly egregious because Venmo's main utility is transmitting money, not being a social network. But making payments public makes the app look juicier and more active to users. 

So, unsurprisingly, the default setting is public. Unless you change your Venmo settings to be private (which you should!), everyone knows who you went out to drinks with.

What to do? Regulation has weaknesses, and can create annoying loophole-exploiting behavior, like we've all seen with cookie-question pop-ups. 

But still, regulation can help. In its AI data-training sweep, LinkedIn excluded countries covered by the European Union's General Data Protection Regulation. And after UK regulators complained, LinkedIn agreed to exclude UK accounts as well.

"Sometimes outrage works," Tobac said. Her firm still advocates for a clear, basic idea: privacy and security should be set by default, not by opt-in.

The big story

A team of almost 200 digital staffers hadn't made much headway when they were working on behalf of US President Joe Biden's reelection campaign. But when Biden dropped out and Vice President Kamala Harris took his place in the race, the staff showed it knew how to make a meme strategy work.

One to watch

Get fully charged

The FCC leveled a $6 million fine against the consultant responsible for AI-generated robocalls against US President Joe Biden earlier this year.

Tokyo Electron prepares for an expansion in India as the government there lures chipmakers.

The US energy secretary says she's open to the idea of tech companies receiving investments from overseas to help build data centers in the US.

More from Bloomberg

Get Bloomberg Tech weeklies in your inbox:

  • Cyber Bulletin for coverage of the shadow world of hackers and cyber-espionage
  • Game On for reporting on the video game business
  • Power On for Apple scoops, consumer tech news and more
  • Screentime for a front-row seat to the collision of Hollywood and Silicon Valley
  • Soundbite for reporting on podcasting, the music industry and audio trends
  • Q&AI for answers to all your questions about AI

Stay updated by saving our new email address

Our email address is changing, which means you'll be receiving this newsletter from noreply@news.bloomberg.com. Here's how to update your contacts to ensure you continue receiving it:

  • Gmail: Open an email from Bloomberg, click the three dots in the top right corner, select "Mark as important."
  • Outlook: Right-click on Bloomberg's email address and select "Add to Outlook Contacts."
  • Apple Mail: Open the email, click on Bloomberg's email address, and select "Add to Contacts" or "Add to VIPs."
  • Yahoo Mail: Open an email from Bloomberg, hover over the email address, click "Add to Contacts."

No comments:

Post a Comment

You’re forgetting your first trade.

Important, your next trade alert. To view this email as a web page, go  here. To view this email as a web page, go  here.               ...