YouTube says it will end full-day masthead ad reservations in 2021, to give “advertisers more budget flexibility”, says change is not connected to the election — - YouTube will discontinue full-day reservations for its masthead ads beginning in 2021.
PayPal reports Q3 revenue of $5.46B, up 25% YoY, as it adds 15.2M net new active accounts; Venmo's total processed payment volume rises 61% YoY to $44B — PayPal said it processed a record-setting $247 billion in total payment volume during the quarter. — PayPal on Monday delivered …
Twitter says, in a new filing, that a board committee formed this spring recommended that current management structure remain in place with Jack Dorsey as CEO — Elliott Management initiated the review of company's leadership in the spring — Jack Dorsey's job as CEO of Twitter Inc. appears safe.
Stensul, which aims to streamline the process of building marketing emails, has raised $16M Series B led by USVP — Stensul, a startup aiming to streamline the process of building marketing emails, has raised $16 million in Series B funding. — When the company raised its $7 million Series …
AWS introduces new GPU-equipped P4 instances, powered by Intel Cascade Lake processors and Nvidia's A100 Tensor Core GPUs, offering 2.5x DL performance over P3 — AWS today announced the launch of its newest GPU-equipped instances. Dubbed P4, these new instances are launching a decade …
The election results will start to come in as early as 7pm Eastern Time on Tuesday, when seven states begin closing the polls. The next few hours will see more polls close around the country, more votes processed, more counts updated. But we won’t have the final result that night.
This isn’t unusual: In the US, counting votes and officially certifying them always goes on longer than Election Day, and the coronavirus means the counting will probably take longer than that. But on Sunday, Axios reported that President Trump intends to prematurely declare victory if it looks like he’s leading in the early returns, even if there are still millions of votes left to be counted. He has denied this specific claim, but it is in line with his long campaign to undermine the legitimacy of the election, and matches his promise to use lawyers to stop ballot counting in Pennsylvania as soon as polls close—even though the state will still have many mail-in ballots left to count and report.
So what exactly will happen if a candidate prematurely declares victory before the contest is truly over?
This the front line. Any premature declaration will likely hit American networks like Twitter, Facebook, and YouTube first, so the way these platforms handle this kind of activity will inform what happens next. Those three sites are planning to use labeling to deal with this kind of disinformation.
Twitter, the president’s social media platform of choice, says it will prominently label misleading tweets about election results from candidates, as well as any viral tweet. Disputed announcements will be met with a label that says “Official sources may not have called the race when this was Tweeted.”
To confirm results, the company will be leaning on state and local election officials as well as major national news outlets with dedicated election coverage desks. At least two sources will have to confirm the results of a race before a candidate can tweet about results without a warning label being applied.
YouTube, which has been a top campaign advertising battleground, will place an information panel on videos prematurely declaring victory. That will link to Google’s election results feature, which is being produced in partnership with the Associated Press.
“We’ll also continue to raise up authoritative content from news organizations and reduce the spread of borderline election misinformation,” said Google spokesperson Ivy Choi. “Additionally, if a piece of content, in the course of prematurely declaring victory, misleads viewers about voting or encourages interference in democratic processes, we will remove that in accordance with our community guidelines.”
When the polls close, all Google’s ad platforms—including YouTube and its search engine—will pause ads that reference the 2020 election. That may cut off another potential avenue for disinformation across the company’s internet empire.
Facebook is placing its own hopes in labels as well, including a preemptive notification in news feeds to follow authoritative news outlets like Reuters and the Associated Press for election results. Facebook’s normal response to false news is to reduce its spread on the network and partner with fact-checkers for additional labeling.
Elsewhere, TikTok’s policy reduces the visibility of posts prematurely claiming victory and is working on an “expedited” schedule with fact checking partners around Election Day.
This is an excerpt from The Outcome, our daily email on election integrity and security. Click here to get regular updates straight to your inbox.
Tomorrow is Election Day in the US, which means we’ve reached peak political saturation: Americans are being hit with constant news alerts, a torrent of punditry and campaign ads on television, and even warring yard signs. The stakes are high, and we’re all struggling to figure out what’s fact and what’s fiction.
Kids and teens are no different. Being young has never been easy, but it’s especially tough when social media, television programs, and maybe even the adults in your life often twist truth into misinformation.
Here are some tips for grownups and young people alike for how to talk with someone about misinformation and make sure the information you’re getting and sharing is true.
We don’t know much about how kids are affected by conspiracy theories and misinformation. “There is so little research examining conspiracy beliefs in younger people,” says Karen Douglas, a professor of social psychology at the University of Kent and the mother of two teenagers. The literature has made clear that more education helps shield people against misinformation, and that same logic probably applies to kids, who may be likelier to believe misinformation the younger they are. Douglas is developing a psychometric scale to measure conspiracy theory belief in adolescents, but until then, we won’t quite know how kids take in misinformation—which makes fighting against it more difficult.
Be age appropriate. Not all kids are ready to handle the graphic details of George Floyd’s murder or the systemic racism underlying it, for example. Nor should they be if they’re younger than tween-age, says Tanner Higgin, the director of education editorial strategy at CommonSense Media. “For kids under seven, don’t involve them in political discussions or worrying about issues,” he says. Younger kids need to know they are safe and parents are keeping them safe, and worrying them—especially during a pandemic, when they have less contact with their friends—will backfire.
That said, don’t sugarcoat. If you’ve got a particularly precocious, mature kid who’s asking pointed questions and can digest information without spiraling into anxiety and worry, be clear and honest. Lying won’t help kids who will undoubtedly find out the truth elsewhere. “Even toddlers can understand how not telling the truth, or basing decisions on bad information, can be harmful,” says Peter Adams, the senior vice president at the News Literacy Project. “They can also understand foundational journalistic concepts like fairness or the importance of accuracy. You just need to tailor the examples or themes you employ to make this real to them.”
Try introducing a “lite” conspiracy theory. This might go against logic, but Douglas says that doing so is important, especially for more gullible little ones: “Once they believe in conspiracy theories, these beliefs are difficult to correct.” Protect against that by introducing a weak version of the misinformation before they’re exposed to it in the real world, and debunking it with them. This helps kids understand what’s problematic about the reasoning, so when a more persuasive conspiracy theory comes along, they’re able to step back and question it.
Remember that you, too, can fall for misinformation. Yes, even you. “A lot of teens—particularly those who are tech savvy—think that they’re too sharp to fall for misinformation so they don’t have to worry about it,” Adams says. But it bears repeating: No one is immune from misinformation.
Be wary of reposts. “If a claim or screenshot is crossposted to a different platform, it could be a sign that it’s missing context,” Alexa Volland says. She knows: Volland trains teen fact checkers across the United States with MediaWise’s Teen Fact Checking Network (a collaboration between the Poynter Institute, the Google News Initiative, and Facebook). She’s seen plenty of Instagram stories featuring screenshots within a screenshot, or screenshots of tweets posted within an Instagram story or TikTok. Solution: Go to the original platform and check out what that person was saying before sharing.
Even toddlers can understand how not telling the truth, or basing decisions on bad information, can be harmful.
Reverse image searches are your best friend for meme checks. On social media, people sometimes post striking images that they think are about a particular news event but actually have nothing to do with that incident. Volland says that doing a simple reverse image search is one of the easiest, fastest ways to check if a viral image is really what it purports to be.
Ask yourself who is behind the information. Look at the organization or person who originally shared the story and think about their possible incentives. What do they stand to gain from the information that was shared? They may be motivated to twist the truth in ways that can lead to misinformation.
Get proof. Be your own fact checker and try to verify the information to the best of your ability. Consider: What is the evidence? Are there links to sources? Are those sources reliable? And are multiple other sources saying the same thing? Sites like FactCheck.org and PolitiFact might be useful here.
Check your own bias. Hello, confirmation bias: If you have a strong reaction of “Ugh, that’s disgusting!” or nod vigorously in agreement with a post, step back. “If it’s a claim that sparks an intense emotional reaction, that can translate into validation,” Volland says. That makes us more likely to believe misinformation.
Check for context. Volland says a lot of the misinformation that goes viral on social media pulls images out of context for memes. For example, her group debunked a viral image supposedly about recent Black Lives Matter protests that misled viewers with images from protests in Ferguson, Missouri, a few years ago.
Go private. Nobody likes being attacked, whether it’s at the dinner table or in the comments section on Facebook. Talk to someone who might be misinformed privately and separately, whether it’s in DMs or in person away from others.
Seek other perspectives. “We tend to read an article up and down but it’s important to open up multiple tabs and get out of your echo chamber,” Volland says. That means going to a news source that might lean the opposite way you tend to, or reading the tweets and press releases of politicians you disagree with. It might be hard, but it will make you more well-rounded and help you know what’s true and what’s inflated.
Check the comments. The comments section will often do a lot of the work in determining whether something is true or not by pointing to alternative sources, and it can be a fast, easy way to see if the post has been flagged by others as suspicious or misleading.
Conversation means meeting the other person halfway. As we’ve said before, being kind is ultimately the most powerful way to talk about misinformation. Attacking people for their beliefs can cause them to double down. Volland suggests “swapping sources” when a news story comes under dispute. Another tip: If a person is skeptical of a news source, presenting information from that source won’t be persuasive. Volland suggests instead seeking out a source you both can agree on and finding information there.
In April, Amazon announced it would contribute $10 million to a pair of projects designed to pay forest owners across the Appalachian Mountains to manage their lands in ways that capture more carbon dioxide from the air.
It is one of the first investments in the retail giant’s $100 million Right Now Climate Fund, an initiative to support the use of “nature-based climate solutions” like forests, grasslands, and wetlands to absorb more of the greenhouse gas. Amazon launched it last year in partnership with the Nature Conservancy, a conservation nonprofit, as part of its effort to reach “net zero carbon” by 2040. Together, they’re developing ways Amazon and other companies can effectively pay others to prevent or remove enough emissions to counterbalance those from their own operations.
But offsets researchers reviewed one of the proposals for the forests in the eastern US on behalf of MIT Technology Review and warned it may significantly overstate carbon reductions. Moreover, studies and articles have repeatedly highlighted similar problems with other offset programs designed to incentivize additional carbon uptake through things like trees and soil.
That’s raising real concerns as a growing list of large companies, including Amazon, Microsoft, and even oil and gas giants like Shell, trumpet “net zero emissions” plans that will rely heavily on nature-based offsets to theoretically cancel out their continuing climate pollution.
Trees suck carbon out of the air through photosynthesis and store it in their trunks, leaves, roots, and branches. Healthy forests with larger trees generally capture more carbon than overpacked forests, where smaller trees and other vegetation compete for water, sunlight, and space. When trees fall down and rot, or are cut down and converted into products like paper, much of the carbon is released back into the atmosphere.
The Nature Conservancy partnered with the American Forest Foundation to create a new offset protocol designed to allow owners of small tracts of wooded land to earn credits for taking steps to suck up and store more carbon.
The Family Forest Impact Foundation, an affiliate of the American Forest Foundation, will pay the landowners for carrying out two practices: promoting the growth of larger trees by harvesting less than previously planned, and thinning out competing shrubs and other vegetation. The change in practices must persist for 20 and 10 years, respectively.
The Family Forest Impact Foundation will, in turn, sell credits for the additional carbon that builds up on the properties to companies like Amazon on voluntary offset markets.
Family landowners largely haven’t participated in such markets up to now because complying with the programs can be complicated and expensive.
“Existing carbon forest markets weren’t working for small landowners,” says Christine Cadigan, director of the Family Forest Carbon Program at the American Forest Foundation. By easing some of the most cumbersome requirements, the groups believe they can reduce the costs by 75%, she says.
The organizations are working with Verra, a nonprofit that accredits offset protocols, to “review and validate” the approach. In the second Amazon-funded effort, known as Forest Carbon Co-ops, the Nature Conservancy is collaborating with the Vermont Land Trust to develop a similar program for owners of wooded lands ranging in area from 200 to 2,000 acres.
Amazon said the two programs together will draw down or prevent the release of 18.5 million metric tons of carbon dioxide by 2031. The company didn’t respond to inquiries from MIT Technology Review before press time.
Several outside researchers who have looked at the proposal, however, fear there are a few ways the program could overestimate the carbon reductions actually achieved.
The biggest red flag for Barbara Haya, a research fellow at the Center for Environmental Public Policy at the University of California, Berkeley, is how the program deals with what’s known as “leakage.” This occurs when reduced timber harvests brought about by offset projects simply lead to increased harvesting elsewhere.
Haya says some earlier research suggests that more than 80% of such reductions can simply shift to harvesting on timberlands in neighboring regions or even other nations. But under the rules for the reduced harvesting practice, landowners would generally only need to account for a 10% leakage rate in their calculations.
This suggests that even if the family forest projects do draw down significant additional carbon, much of the benefit could be wiped out by larger harvests elsewhere, limiting the real-world climate benefits.
Some observers also worry about how the projects will be audited to ensure compliance.
One of the key ways the program promises to make participation less expensive is by eliminating the need for surveyors to come out and conduct detailed assessments of every project site.
Instead, the program will use an aggregation of sample plots in similar forests to figure out what would be expected to happen on the project land in the absence of the program, given common forestry practices in the region. They’ll then compare those figures with field measurements of additional stored carbon over time from a “statistically significant random sample of properties” enrolled in the program, to determine how much more carbon the practices should be saving or removing.
This approach may produce an accurate accounting over time, says Grayson Badgley, a plant physiologist at Black Rock Forest and Columbia University. But he says it will be tricky to ensure that all the assumptions are correct, and that they properly select and weight plots to reflect conditions and land management practices on the enrolled projects.
One risk that is that the forestry practices assumed to be common in the area could be more representative of large timber companies than family landowners. That would exaggerate the amount of harvesting that would have occurred in the program’s absence, thus overstating the carbon gains it achieves.
Finally, there are additional concerns about whether the program will bank enough credits to account for setbacks that could occur if landowners simply increase harvesting at the end of the 10- or 20-year contract terms—or as a result of natural risks to trees like wildfires, storms, and insect infestations, all of which are rising with climate change.
In an email, Cadigan stressed that they’re still in the approval process and are working through various adjustments based on public comments and other feedback. But she also said they’re confident that their methodologies will bring about sustained improvement in forestry practices and accurately estimate additional carbon removal over time.
“Once they’ve reset their management, it actually makes more economic sense for them to maintain this approach, and as a result, this management will have a long-term positive impact,” she wrote.
The family forest program is just one of numerous offset efforts that Amazon intends to eventually invest in or purchase credits from. The company also announced plans to provide more than $4 million to an “urban greening” program in Germany, another Nature Conservancy project.
Amazon is taking concrete steps to cut its direct corporate emissions as well. It’s invested in more than 30 large-scale solar and wind projects around the world and added rooftop solar panels to dozens of fulfillment or sorting centers, as part of its effort to run entirely on renewable electricity by 2025. The retailer also agreed to purchase 100,000 electric delivery vans from Rivian with an eye toward ensuring that half of its shipments are “net zero carbon” by 2030.
But between its corporate facilities, data centers, operations, and suppliers, the company has a massive carbon footprint – and one still growing at last count. Last year it emitted the equivalent of more than 50 million metric tons of carbon dioxide, directly or indirectly. That’s up from around 44 million in 2018.
Amazon, like most companies, hasn’t specified what portion of its emissions it expects to address through nature-based offsets. But a heavy reliance on them creates very real challenges if most of these programs are, as a growing number of researchers believe, often overcounting actual reductions.
It allows companies to assert to customers, policymakers and others that they’re operating in a climate neutral way, while continuing to produce more planet-warming gases, on a ton-for-ton basis, than the programs are removing.
Another issue is that the growing number of nature-based projects is creating larger pools of low-cost carbon offsets, the availability of which can undermine the viability of more reliable carbon-capture methods.
Bottom-line-minded companies will, for instance, likely pick a roughly $10 forestry offset that purports to cancel out the same ton of emissions that Swiss startup Climeworks is charging $1,100 to reliably remove and permanently store, using carbon dioxide sucking machines and underground geological formations. (Notably, Microsoft has said it only wants to pay $20 a ton for offsets as it looks to cancel out its entire corporate history of emissions, which some observers believe will steer it away from the more dependable means of carbon removal.)
It will also often be far cheaper for a corporation like Amazon to buy offset credits than to figure out the tougher aspects of corporate emissions reductions, like fully cleaning up the shipping process or ensuring that its vast network of suppliers is carbon free.
“You are essentially giving these large corporations a license to continue doing business as usual,” says Sam Davis, a conservation scientist at the Dogwood Alliance, an environmental nonprofit focused on protecting forests in the southern US. “If we really need and want to address climate change from a corporate perspective, then we can’t just pay the debt with fancy carbon credits and greenwashing.”
Climate models show that the world will now need to slash emissions and draw down billions of tons of carbon dioxide per year by midcentury to prevent really dangerous levels of global warming. But there are limits to how much forests and other nature-based systems can do to get us there.
Ideally, these options should be reserved for the really hard-to-solve parts of the decarbonization puzzle—like aviation, heavy industry, and methane from agriculture—or used to grant poor nations leeway to continue emitting a little longer as their economies develop, says Holly Buck, an assistant professor of environment and sustainability at the University of Buffalo.
In other words, there are real risks if rich companies in rich nations buy up a disproportionate share of the cheapest sources of carbon removal while they’ve got plenty of other ways to drive their emissions toward zero.
Due to Indonesia’s vast size and population, timely and reliable access to healthcare can sometimes be a challenge. Halodoc aims to change that with a mobile first-telemedicine platform that connects Indonesians to doctors and helps them arrange appointments, medicine deliveries and tests.
What’s distinctive about the Halodoc platform is that it draws on human-centered artificial intelligence: a promising new area of research that uses continuous human feedback to improve how AI systems work, and provides a better experience for the people who rely on those systems.
With support from Google’s Late Stage Accelerator, a program that assists high-potential startups, we assembled a team of doctors, data scientists, engineers, product managers and researchers to determine how technology could support Indonesian doctors’ work. One particular approach the team identified was using AI to replicate the mentoring and feedback that junior doctors receive from more experienced colleagues in hospitals—a process that’s important to improving quality of care, but is hard to reproduce on a larger scale.
We set out to create an easy way to provide feedback in virtual health, and worked with Google’s machine learning experts in the Late-Stage Accelerator to determine the best approach. With Google’s guidance, Halodoc's engineers applied Natural Language Processing in Bahasa Indonesia to measure, rank, and provide insights that can inform doctors’ decisions across the country—using thousands of consultations to train their machine learning models.
When doctors open the Halodoc app, they see information on how they performed based on their response time and quality index metrics, along with suggested actions on how they can improve their consultation quality. They also have the option of receiving further feedback and coaching from more senior doctors if needed.
Right now, more than five percent of Indonesians use Halodoc’s platform. As a result of applying AI principles to improve the quality of care that patients experience, our app ratings have increased from 4.5 to 4.8 stars in fewer than six months, while our overall doctor scores have improved by 64 percent.
Halodoc’s telemedicine app enables doctors to deliver personalized feedback with assistance from ML-enabled insights that improve patient care.
From here, with Google’s help, we hope to continue simplifying Indonesia’s healthcare infrastructure and advance the application of AI in healthcare globally.