Arunesh Mathur

In a new paper, @jonathanmayer, Mihir Kshirsagar, and I investigate a question that has been challenging researchers and policymakers: what makes a dark pattern, well, "dark"? arxiv.org/pdf/2101.04843… [thread]

We’re inviting comments on the partnerships, and other arrangements, between #Microsoft and #Mistral #AI, #Amazon and #Anthropic, and Microsoft and #Inflection AI. Read more: gov.uk/government/new…

@LauraEdelson2 I think you're being way too generous in entertaining this argument. Not being able to charge for something isn't a sufficient criterion to describe an anticompetitive marketplace. :-)

This has been going on for a while and has cycled through a few justifications. Previous ones have been a bit silly (China! SkyNet! Biotoxins!) but the current one (Open is Anticompetitive!) isn't quite as ridiculous on its face, so I want to explain why it's dead wrong.4/12

I'm so tired of being in rooms where people whisper about the absolute ARMY of Big Tech-funded people (most, but not all, ex-Googlers) that have popped up in nearly every corridor in DC where people are working on literally anything to do with AI. So let's talk about it! 1/12

We’ve now published our technical update report following last week’s update on #AI foundation models. Read more: gov.uk/government/pub…

We’ve outlined our growing concerns in the #AI #foundationmodels market, including 3 key interlinked risks to fair, open and effective competition. Read more about these, as well as our plans for further action, here: gov.uk/government/new…

Today, our CEO, Sarah Cardell, has launched our initial report into #AI Foundation Models, which proposes new principles to support competition and protect consumers. Find out more: gov.uk/government/pub…

@josephfcox For reference, @TfL allows this too but 1) you need to create an account on their website, and 2) they run an active card check so your card provider tells you that the card has been used on their website. It's not a perfect system either but provides some level of transparency.

New: I tracked the precise movements of an NYC subway rider. Saw what specific time they got on and at what station. It became obvious which station was nearest to their home. This was all because of a 'feature' on the MTA website Wide open to abuse 404media.co/i-tracked-nyc-…

Community Notes may have been originally designed to identify misleading tweets but the fact that they work on ads as well has people pushing back against businesses that aim to deceive.

@hoofnagle @Susan_Athey "Post-shutdown, they read less breaking news, hard news, and news that is not well covered on their favorite news publishers."

@hoofnagle @Susan_Athey "We find that the shutdown of Google News reduces overall news consumption by about 20% for treatment users, and it reduces page views on publishers other than Google News by 10%."

I suspect this is good news and that it sets up a natural experiment to see whether Facebook as a news source is correlated with disinformation and extremism arstechnica.com/tech-policy/20…

We’ve launched an initial review of artificial intelligence models: gov.uk/government/new…

Thanks Bing, but unfortunately, I will have to pass.

@BedoyaFTC @DavidDorrell1 @CMAgovUK @DavidDorrell1 got to it before I could but yes the @CMAgovUK has a few in-house behavioral scientists that do terrific work.

@DavidDorrell1 @CMAgovUK Thank you!

Is anyone aware of any consumer protection agency (here or abroad) with oversight over tech companies which has a full-time psychologist on staff -- not just part-time or as an external contractor? I am not aware of any, but would love to know if I've missed something.

@ProfFionasm These are excellent—thank you! I'd love to see more content that covers digital markets. I'm thinking both along the lines of fundamental issues (where they might differ) as well as remedies and upcoming regulation. Perhaps drawing from case studies too?

I'll be presenting our paper on dark patterns at 7pm EDT tonight and 11am EDT tomorrow. Come learn how HCI researchers can collect evidence that proves when a design pattern might become "dark". #CHI2021 #darkpatterns

Here's what many political campaigns and organizations are upto: twitter.com/aruneshmathur/…

This seems like a good time to remind everyone that #darkpatterns such as these are not just an artifact of the Trump campaign. They're used across the political spectrum to different ends. We've been documenting their types using our corpus here: electionemails2020.org.

@uma_karma @knutson_brain @R_Thaler We argue that these normative lenses provide the means to judge designs as dark patterns. Of course we still need to determine when each of these is "sufficiently" diminished to make this call. The paper provides an extensive discussion about this in Section 5.

@uma_karma @knutson_brain @R_Thaler We specify several "negative/evil" outcomes in the paper: 1) diminishing individual welfare (e.g., financial loss, loss of privacy), 2) diminishing collective welfare (e.g., market competition, trust), 3) undermining laws, and 4) harming individual autonomy.