and jack dorsey testified before a subcommittee virtual hearing on combating online information and misinformation. you can watch the whole hearing later tonight or anytime on c-span.org. >> good morning. we are californian, so it is good morning for us. i want to start by saying content moderation, like removing posts or banning accounts is about treating symptoms and i think we need to treat symptoms but we also need to address underlying diseases. your products amplify extremism. your business model of targeted ads enables misinformation to thrive because you chase user

engagement at great cost to our society. so last month, the anti-defamation league found that youtube amplifies extremism. scores of journalists and researchers agree. a user watching an extremist video is often recommended more such videos, slowly radicalizing the user. youtube is not doing enough to adjust recommendations and it is why representatives introduced the protecting americans from dangerous algorithms act, that narrowly a men's section -- amends section 230 to curb violence. it is also why i, along with 40 of my house colleagues, wrote to

you. i ask that those be placed into the record. so my question to you is are you willing to overhaul youtube's core recommendation engine to correct this issue? yes or no? >> congresswoman, we have overhauled our recommendation systems, and i know you have engaged on these issues before, pretty substantially. >> yes or no? we still have a huge problem. are you saying that the anti-defamation league does not know what they are talking about westmark that all of these journalists and researchers, there is a lot more to address. that's why i am asking you if you are willing to overhaul youtube's core recommendation engine to correct it? it is serious, dangerous, what

more can i say? yes or no? >> if i may explain -- >> i don't have time to explain. let me just say this to the witnesses. we don't do filibuster in the house. that is something that is done in the senate. filibuster does not work with us. mr. zuckerberg, your algorithms use unseemly amounts of data to keep users on your platform because that leads to mourad revenue. -- more ad revenue. businesses make money, we know that, but your business model produces anxiety and deadly misinformation. the center for countering digital hate found that the explorer suggested posts part of

instagram are littered with covid information, election misinformation and qanon posts. that's why this bill will ban the business model of surveillance advertising. are you willing to redesign your products to eliminate focus on addicting users at all costs? >> as i said before --. >> let me just say this, and i think it is here retaining all of us. that is that no one seems to know the word yes or the word no. which one is it? if you don't want to answer, just say you don't want to answer. >> congresswoman, these are nuanced issues. >> mr. dorsey, i think that you

need to eliminate all covid misinformation and not reduce its spread, but remove it. i looked at a tweet this morning. robert kennedy junior thinks the death of baseball legend hank aaron to the covid-19 vaccine. even though fact checkers debunked the story, the tweet has 19,000 retweets. would you take this down, why haven't you, and would you band the 12 accounts -- ban the 12 accounts spreading deadly covid misinformation? >> we won't take it down because it did not violate our policy. speaker pelosi: what kind of policy is that? is it a policy for misinformation? >> no.

>> the gentlelady's time has expired. the chair recognizes mr. scully's. -- mr. scalise. >> thank you, mr. chairman. i want to thank our three witnesses. we are seeing a lot of concern being expressed by members of both sides about the way these platforms are run and as it relates to the treatment of people. i have had a lot of concerns shared with you individually over the past few years, whether it is algorithms designed to have a bias against conservatives. we all agree that illegal activity and bullying ought not be permeated through social media, but there is a big difference between stopping

bullying and violent types of social media posts and censorship of political views that you do not agree with. let me ask my first question to mr. dorsey. there have been a lot of questions about that unequal treatment and i will start with the new york post article. a lot of people have seen this. this article was censored by twitter when it originally went out. this newspaper goes out to 1801 founded by alexander hamilton, and this incredibly sourced article about hunter biden was banned by twitter. you have this washington post article where parts of it have been debunked, and yet this article can still be tweeted

out. i want to ask mr. dorsey, do you recognize there is a concern there is an anti-conservative bias on twitter's behalf and you recognize this has to stop? twitter is going to be viewed by both sides as a place will get good treatment? >> we made a mistake with the new york post. it does not do with content but the hacked materials policy. we do not write policy according to any particular political leaning. if we find any, we wrapped it out. -- rat it out. we make mistakes. our goal is to correct it as quickly as possible, and in that case we did. >> the new york post twitter account was blocked for weeks.

we have got a first amendment. it seems like a new york -- a newspaper founded in 1801 by alexander hamilton, to be blocked by mistake seems like a big mistake. was anybody held accountable in your censoring department for that mistake? >> we don't have a censoring department. >> who made the decision to block their account for two weeks? >> we don't require them -- we don't block their account. for two weeks we require them to delete the tweet. -- don't block their account for two weeks. we require them to delete the tweet. speaker pelosi: there is going to be more conversation about it, but you are acting as a publisher. if you tell in newspaper they have to delete something to participate, don't you recognize

you are no longer hosting a town square but acting as a publisher? >> it is literally just a process error. this is not against them in any way. if we remove a violation, we require people to correct it. we change that. -- changed that. i see it, it is something we learned. >> let me go back to the washington post article. this can still be tweeted. it contains false information, and yet there are tweet today on your service that still mischaracterize it in a way were even the washington post admitted it is wrong yet those mischaracterizations can still be retweeted. will you correct that? >> our misleading information policies are focused on

manipulating media, public health, and civic integrity. >> i would hope you would take that down. i know you said you are running a business and a business quote wants to grow. i want you to recognize that if you continue to be viewed as an anti-conservative platform, there will be other people that step up to compete. i hope you recognize that. >> the gentlemen's time has expired. the chair now recognizes mr. butterfield. >> thank you mr. chairman. last year, in response to the police killing of george floyd, you wrote a post on your facebook page that denounced racial bias. it proclaimed black lives matter. it announced the company would donate to racial justice organizations.

and mr. dorsey, twitter changed its official bio to a black lives matter tribute and you pledged millions to an organization started by colin kaepernick. google held a moment of silence to honor george floyd and you announced $12 million in grants to racial justice organizations. the ceo of youtube wrote in a blog post quote we believe black lives matter and we all need to do more to dismantle systemic racism. due to announced it would start a $100 million fund for black creators. all of this sounds nice these pronouncements and donations do not address the way your company's own products. the way you are being weaponized by racists, being weaponized

against commuters of color and spreading racist content. in my view, your companies have contributed to the spread of race-based extremism and voter suppression. the head of mcdonald's -- it would be like if the head of mcdonald's, burger king, and taco bell got together to lower obesity. you could have made meaningful changes in your organization to address the biases built into your products, but instead we are left with platitudes and another round of passing the buck. america is watching you today. this is a moment that begins a transformation of the way you do business and you must understand that. perhaps a lack of diversity in your organization has contributed. the black caucus 2025 initiative

has been working for years to increase diversity and equity tech companies at all levels, and you know that because we have visited you in california. we founded in 2015, with the hope that by now, the tech workforce would reflect the diversity of our country. i acknowledge you have made some modest advancement, but not enough. must be meaningful representation in your companies to design your products and services in ways that work for all americans. that requires public accountability. history has shown that you have talked the talk but failed to walk the walk. it appears now that congress will have to compel you to make meaningful changes. i will strive for a yes or no answer and hopefully have better results than my colleagues. mr. zuckerberg, yes or no, would

you oppose legislation that would require technology companies to publicly report on workforce diversity at all levels. >> i don't think so, but i need to understand it in more detail. >> we will talk about that and i hope you will not oppose. what about you, mr. dorsey? would you oppose a law that would make workforce reporting a requirement? >> i would not oppose it. it would come with some publications. -- complications. >> thank you. what about you, mr. fitch i -- mr.pichi -- mr.pichai? >> we were the first company to publish transparency reports.

we publish it annually so we are happy to share that with you and take any feedback. what we do provides detailed demographic information and we can always do better. >> the congressional black caucus has said to you over and over again, we need greater diversity among your workforce and we need for you to publish the data so that the world can see it. thank you mr. chairman. i yield back the 10 seconds that i have. >> the gentleman reserves commendation for doing that and i hope others follow his example. the chair now recognizes mr. guthrie. >> thank you mr. chair, and thank you to the witnesses. >> thank you, mr. chair, and thanks for the witnesses for being here. big tech decisions have real impact on people. that's why i ask my constituents using those platforms to share their experiences with me. and i'm here to advocate on their behalf. i received 450 responses.

one major thing i heard was the experience websites taking down religious content, which is important because there's a lot of religious organizations now streaming their services due to covid. of religious organizations now streaming their services due to covid. i did have some one stance where a constituent wrote to me, quote, and this is what she posted. i am thankful god's grace is new every morning, and then facebook took it down, and my constituent said she got a notice from facebook that it violated their policies around hate. and so i just want to discuss about this. i could ask you yes or no questions, mr. zuckerberg, on that. but i just want to talk it a little bit. i know that we don't want extreme language on the internet. i'm with you on that. and you can't watch everything. so you use algorithms to find that. so algorithms will flag things, some that are clearly obvious

and some that probably shouldn't be flagged. but it seems to me that it seems to be bias in that direction. instead of just giving a yes-or-no question, i'm going to read that question. within that quote, what in there would get tripped up and would this quote get tripped up and put into the flagged category. it says, i am thankful god's grace is new every morning. and so i guess the question is what word or thought do you think would trip an algorithm for that quote, mr. zuckerberg? >> the congressman, it is not clear to me why that post would be a problem. i would need to look into it in more detail. sometimes the systems look at patterns of posting. so if someone is posting a lot, then maybe our system thinks it's spam. but i would need to look into it in more detail. overall the reality is that any system is going to make mistakes. there's going to be content that we take down that we should've left up. and there's going to be content

that we missed that we should've taken down that we didn't catch or that the systems made a mistake on. and at scale unfortunately those mistakes can be a large number even if it's a very small percent. but i think that that's why when we're talking about things like section 230 reform, i think it is reasonable to expect large companies to have effective moderation systems but not reasonable to expect that there are never any errors. but that i think that transparency could help hold the companies accountable as to what accuracy and effectiveness they're achieving. >> okay. i think they did receive a notification it was for the hate policy. so -- and i understand there's going to be gray areas, whatever. but that quote, i don't see where the gray areas and how it could get caught up in that. >> i agree. >> thanks for your answer with that. i want to move on. so, mr. dorsey, i want to talk

about the -- i didn't see that quote, but you said that didn't violate your policy. and in the context of that, i know cdc just recently updated its school guidance to make clear science says you can be three feet away and says you can be safe in schools. things are changing every day because we're learning more and more about this virus. how did that not violate your policy, rfk jr., and we have an rfk and jfk iii. but rfk jr., and the policy towards that and then how do you keep up with what's changing so quickly, mr. dorsey? >> you know, we can follow up with you on the exact reasoning. but we have to recognize that our policies evolve constantly and they have to evolve constantly. as has been said earlier in this testimony, we observe what's happening as a result of our policy. we've got to understand the ramifications and we improve it.

and it's a constant cycle. we're always looking to improve our enforcement. >> so, mr. zuckerberg, mr. pichai, just on that continuously evolving information on covid, because we are learning more and more about it, and how do you keep up, only about 30 seconds, so if you could quick answer for each of you, if you can. mr. pichai, maybe. >> on covid, we've been really taking guidance from cdc and other health experts proactively. one thing we get to do in youtube is recommend higher quality content. we have show 400 million information panels last year including a lot from cdc and other health organizations. >> okay. thank you. and i yield back four seconds, mr. chair. >> thank you, mr. guthrie. chair now recognizes ms. matsui for five minutes. >> thank you very much, mr. chairman, for having this hearing today.

today we have another opportunity here from the leaders of facebook, twitter, and google. and one has become a concerning pattern. the members of this committee are here to demand answers to questions about social media's role in escalating misinformation, extremism, and violence. last week i testified at a house judiciary committee hearing about the rise in discrimination and violence against asian-americans. that hearing came on the heels of a violent attack in atlanta that left eight people, six of whom were asian women, dead. the issues we are discussing here are not abstract. they have real world consequences and implications that are too often measured in human lives. i'm worried, as are many watching this hearing, that the companies before us today are not doing enough to prevent the spread of hate, especially when it's targeted against minority communities. clearly the current approach is not working, and i think congress must revisit section 230.

a recent study from the university of san francisco examined nearly 700,000 tweets in the week before and after president trump tweeted the phrase "chinese virus." the results show two alarming trends. there was a significantly greater increase in hate speech the week after the president's tweet, and that half of the tweets used in the hashtag china virus showed an anti-asian sentiment compared to just one-fifth of the tweets using the hashtag covid-19. this evidence backs up what the world health organization already knew in 2015, saying disease names really do matter, we've seen certain disease names provoke a backlash against members of particularly religious or ethnic communities. despite this, facebook and twitter are still allowing hashtags like china virus, kung flu, and wuhan virus to spread. mr. zuckerberg, and mr. dorsey, given the clear association between this type of language

and racism and violence, why do you still allow these hashtags on your platforms? anyone want to answer that, or is that not answerable? >> i think we are waiting for you to call on one of us. we do have policies against hateful conduct, and that includes trends. so when we see associated with any hateful conduct, we will take action on it. it's useful to remember that a lot of these hashtags do contain counterspeech. and people on the other side of it do own them and show why this is so terrible and why it needs to stop. >> can i just take my time back? the fact of the matter is -- algorithms to kind of get rid of these things. mr. zuckerberg, any comment here? >> thanks, congress woman. the rise in anti-asian hate is a really big issue and something

that i do think we need to be proactive about. i agree with the comments that jack made on this. on facebook, any of that context, if it's combined with something that's clearly hateful, we will take that down, it violates the hate speech policy. but one of the nuances that jack highlighted that we certainly see as well in enforcing hate speech policies is that we need to be clear about when someone is saying something because they're using it in a hateful way versus when they're denouncing it. and this is one of the things that has made it more difficult to -- >> an opportunity to really look at hate speech and what it really means, particularly in this day and age when we have many instances of these things happening. and hate speech on social media can be baked down. and unfortunately this also is a trend that maybe happened years and years ago, which it might've just been a latent situation. but social media, it travels all

around the world and hurts a lot of people. and my feeling, and i believe a lot of other people's feeling is that we really have to look at how we define hate speech. you all are very brilliant people, and you hire brilliant people. i would think that there is a way for you to examine this further and take it one step lower to see if it is something that is legitimate or not. and i really feel that this is a time especially now in examining platforms and what you can do and should do. and as we're examining here in this committee and as we write legislation, we really want to have the entire >> you can watch this hearing on combating online misinformation with mark zuckerberg, yu darvish i, and jack dorsey tonight on c-span two or anytime online at

c-span.org. >> c-span's washington journal, every day, we take your calls live on the air on the news of the day and we discussed policy issues that impact you. friday morning, we are taking your calls and getting your reaction to president biden's first press conference, then a reporter discusses big tech leaders' testimony before congress. watch washington journal live at 7:00 eastern friday morning and join the discussion with your calls, comments, texts, and tweets. >> we are joined by the chair and ranking member of the rules committee, representative jim mcgovern from massachusetts, and also, from oklahoma, representative tom cole. congressman, thank you for joining us. >> thank you. congressman, thankr joining us. both: thank you.