Category: BUYSEMPERFI

Marwa Fatafta, MENA policy and advocacy director at the nonprofit Access Now, a digital rights advocacy group, says that she has seen little change in Meta’s systems from 2021, and believes that the company’s content moderation policies still lack transparency for users. For instance, following Russia’s invasion of Ukraine, Meta made an exception to its policies around incitement to violence, allowing users in certain countries to post content, such as “death to the Russian invaders”, referring to the Russian military, that would normally be removed. “​​We still won’t allow credible calls for violence against Russian civilians,” Meta spokesperson Andy Stone told the Verge at the time.

“It’s not clear why some of these exceptions are made for some conflicts and not others,” says Fatafta. “We’re seeing videos and photos, sometimes just from bystanders or journalists, being removed and it’s not clear why. We’re really advocating for more context-specific content moderation.”

For some of these, like the post from al-Shifa, the company will assess whether the post is “newsworthy”, and reinstate images or videos when users appeal a takedown decision. This happens “pretty much in every single crisis,” according to Diya. In the case of the hostage video, the user posted it with a caption encouraging people to watch it to gain a “deeper understanding” of what happened on October 7, violating Meta’s long standing policy of not showing terrorist attacks, and its new policy of showing identifiable images of hostages. (Meta temporarily updated its policies to take down videos in which hostages were identifiable after October 7th).

In a company blog, Meta said the “Oversight Board’s guidance in these cases, along with feedback from other experts, will help us to continue to evolve our policies and response to the ongoing Israel-Hamas War.”

But the bigger issue, Diya says, is that the company continues treating each conflict like a one-off situation that requires a tailored response. “There’s a general reluctance within platforms to preempt or prepare for crises, especially if it’s outside the U.S., even when there is a prolonged history of conflict or violence in that region,” she says. “But we have seen enough crises in the last decade to get some sense of some patterns and what kind of tools should be in place.”

The expedited decisions from the Oversight Board, expected within 30 days, may finally push the company to do just that.

Hoffman and others said that there’s no need to pause development of AI. He called that drastic measure, for which some AI researchers have petitioned, foolish and destructive. Hoffman identified himself as a rational “accelerationist”—someone who knows to slow down when driving around a corner but that, presumably, is happy to speed up when the road ahead is clear. “I recommend everyone come join us in the optimist club, not because it’s utopia and everything works out just fine, but because it can be part of an amazing solution,” he said. “That’s what we’re trying to build towards.”

Mitchell and Buolamwini, who is artist-in-chief and president of the AI harms advocacy group Algorithmic Justice League, said that relying on company promises to mitigate bias and misuse of AI would not be enough. In their view, governments must make clear that AI systems cannot undermine people’s rights to fair treatment or humanity. “Those who stand to be exploited or extorted, even exterminated” need to be protected, Buolamwini said, adding that systems like lethal drones should be stopped. “We’re already in a world where AI is dangerous,” she said. “We have AI as the angels of death.”

Applications such as weaponry are far from OpenAI’s core focus on aiding coders, writers, and other professionals. The company’s tools by their terms cannot be used in military and warfare—although OpenAI’s primary backer and enthusiastic customer Microsoft has a sizable business with the US military. But Buolamwini suggested that companies developing business applications deserve no less scrutiny. As AI takes over mundane tasks such as composition, companies must be ready to reckon with the social consequences of a world that may offer workers fewer meaningful opportunities to learn the basics of a job that it may turn out are vital to becoming highly skilled. “What does it mean to go through that process of creation, finding the right word, figuring out how to express yourself, and learning something in the struggle to do it?” she said.

Fei-Fei Li, a Stanford University computer scientist who runs the school’s Institute for Human-Centered Artificial Intelligence, said the AI community has to be focused on its impacts on people, all the way from individual dignity to large societies. “I should start a new club called the techno-humanist,” she said. “It’s too simple to say, ‘Do you want to accelerate or decelerate?’ We should talk about where we want to accelerate, and where we should slow down.”

Li is one of the modern AI pioneers, having developed the computer vision system known as ImageNet. Would OpenAI want a seemingly balanced voice like hers on its new board? OpenAI board chair Bret Taylor did not respond to a request to comment. But if the opportunity arose, Li said, “I will carefully consider that.”

Gamers have long been stigmatized as lonely weirdos. Some of that has been deserved—ask anyone who’s had a bunch of children shout horrible slurs at them during a match of Call Of Duty. But some leaders in the gaming industry want to push back against that narrative by creating games that encourage users to form communities as part of the gameplay. The idea is that by fostering more human interactions, games can promote positivity and openness, bringing people together instead of pushing them apart.

“Whether it’s a town hall meeting for a community or whether it’s a group of gamers getting together in a park, whenever people meet face-to-face, there’s a level of civility, courtesy, and respect that you often see,” says John Hanke, the founder and CEO of Niantic, the developer behind the massively popular augmented-reality mobile game Pokémon Go. He says a big part of cultivating that sort of positive interaction involves designing a game that entices players outside their comfort zones—or, in the case of an AR game like Pokémon Go, actually getting them outside. “It’s just sort of wired into us to be more open to real human contact and not be as quick to withdraw and as heated and nasty as online.”

Hanke’s remarks were part of a panel at LiveWIRED, an event held yesterday in San Francisco for WIRED’s 30th anniversary. The session, called “Will Games Eat the World?”, featured Hanke; Rachel Kowert, the research director at Take This, a nonprofit that cultivates mental health resources for gamers and game developers; and Jade Raymond, the president and founder of Haven Studios, a game developer that was acquired by Sony last year. The panel was moderated by WIRED special projects editor, Alan Henry.

(L-R) Alan Henry, Special Projects Editor at WIRED, John Hanke, Dr. Rachel Kowert, and Jade Raymond speak onstage during Will Games Eat the World? at LiveWIRED on December 05, 2023 in San Francisco, California.Photograph: Kimberly White/Getty Images

Google just launched its Gemini AI model. Want to try it out for free? A version of the model, called Gemini Pro, is available inside of the Bard chatbot right now. Also, anyone with a Pixel 8 Pro can use a version of Gemini in their AI-suggested text replies with WhatsApp now and with Gboard in the future.

Only a sliver of Gemini is currently available. Future releases are expected to include multimodal capabilities, where a chatbot processes multiple forms of input and produces outputs in different ways, just the text-based version has been added to Bard.

Gemini is also only available in English, though Google plans to roll out support for other languages soon. As with previous generative AI updates from Google, Gemini is also not available in the European Union—for now.

Despite the premium-sounding name, the Gemini Pro update for Bard is free to use. With ChatGPT, you can access the older AI models for free as well, but you pay a monthly subscription to access the most recent model, GPT-4. Details on future plans for Gemini remain scarce, Google teased that its further improved model, Gemini Ultra, may arrive in 2024 and could initially be available inside an upgraded chatbot called Bard Advanced. No subscription plan has been announced yet, but for comparison, a monthly subscription to ChatGPT Plus with GPT-4 costs $20.

GeminiCourtesy of Google

How to Access Google’s Gemini Pro

Do you already have a Google account? Using Gemini inside of Bard is as simple as visiting the website in your browser and logging in. Google does not allow access to Bard if you are not willing to create an account. Users of Google Workspace accounts may need to switch over to their personal email account to try Gemini.

Remember that all of this is technically an experiment for now and you might see some software glitches in your chatbot responses. One of the current strengths of Bard is its integrations with other Google services, when it actually works. Tag @Gmail in your prompt, for example, to have the chatbot summarize your daily messages or tag @YouTube to explore topics with videos. Our previous tests of the Bard chatbot showed potential for these integrations, but there are still plenty of kinks to be worked out.

So how is the anticipated Gemini Ultra different from the currently available Gemini Pro model? According to Google, Ultra is its “most capable mode” and is designed to handle complex tasks across text, images, audio, video, and code. The smaller version of the AI model, that’s fitted to work as part of smartphone features, is called Gemini Nano, and it’s available now in the Pixel 8 Pro for WhatsApp replies.

As you experiment with Gemini Pro in Bard, keep in mind many of the things you likely already know about chatbots, such as their reputation for lying. Not sure where to even start with your prompts? Check out our guide to crafting better prompts for Google’s Bard.

It’s been a monumental year for Crispr, the molecular tool scientists use to edit genetic material. This November, the United Kingdom authorized the first medical treatment using Crispr gene editing, giving people with sickle cell disease new opportunities to receive a one-time therapy to prevent episodes of terrible pain. This week, the US Food and Drug Administration is poised to make a decision about the therapy. What was once seen as a moonshot is already changing lives.

Right now, though, it’s still a rarefied treatment. “It’s expensive,” Jennifer Doudna, the pioneering biochemist who won a Nobel Prize in 2020 for her work on Crispr, told WIRED’s Emily Mullin at the LiveWIRED conference this week in San Francisco. The therapy is expected to be priced at over a million dollars a patient, which could make it inaccessible to many of the people who need it most.

It’s also a complicated process; patients have stem cells taken from their bodies, edited in laboratory settings, and then put back in. Doudna is optimistic for a future where Crispr-based treatments are far less invasive than they are now. “Maybe even a pill at some point,” she says. “Today that sounds a little bit fantastical, but I think it’s very achievable.”

In 2014, Doudna founded the Innovative Genomics Institute to apply Crispr technology to healthcare questions. Doudna hopes that the IGI’s research can also help make these technologies more affordable and accessible; she’s also very interested in how Crispr might be used to fine-tune the microbiome.

Emily Mullin, Staff Writer at WIRED, and Jennifer Doudna speak onstage during The New Age of Medicine at LiveWIRED 2023.Photograph: Kimberly White/Getty Images

One attraction of Binance, as the company grew from its 2017 founding into the biggest cryptocurrency exchange in the world, was the firm’s freewheeling flouting of rules. As it amassed well over 100 million crypto-trading users globally, it openly told the United States government that, as an offshore operation, it didn’t have to comply with the country’s financial regulations and money-laundering laws.

Then, late last month, those years of brushing off US regulators caught up with the company in the form of one the most punitive money-laundering criminal settlements in the history of the US Justice Department. The crackdown doesn’t just mean a chastened Binance will have to change its practices going forward. It means that when the company is sentenced in a matter of months, it will be forced to open its past books to regulators, too. What was once a haven for anarchic crypto commerce is about to be transformed into the opposite: perhaps the most fed-friendly business in the cryptocurrency industry, retroactively offering more than a half-decade of users’ transaction records to US regulators and law enforcement.

When the Department of Justice announced on November 21 that Binance’s executives had agreed to plead guilty to criminal money-laundering charges, much of the attention on that settlement focused on founder Changpeng Zhao giving up his CEO role and on the company’s record-breaking $4.3 billion fine. But Binance’s settlement agreements with the DOJ and the US Treasury Department also stipulate a strict new regime of data-sharing with law enforcement and regulators. The company has agreed to comply with regulators’ “requests for information”—a term that carries none of the evidence or suspicion requirements necessary for obtaining a warrant or even a subpoena—to the point of producing any “information, testimony, document, record, or other tangible evidence.”

Binance has also agreed to scour all of its transactions from 2018 to 2022 and file suspicious activity reports (SARs) for anything it deems a potential violation of US law from that five-year period. That “SAR lookback” means the company will now be actively scrutinizing its customers in retrospect, not just passively assenting to regulators poring over its databases. Those SARs are collected by FinCEN, the Treasury Department’s financial crimes division, but then made available to law enforcement agencies from the FBI to IRS Criminal Investigations to local police. And all of this new scrutiny will be overseen by a “monitor” firm chosen by the US government but paid by Binance—an in-house watchdog assigned to make sure Binance is complying in good faith.

“I don’t think Binance’s customers have the slightest clue of the ramifications of this plea and consent decree. It’s unprecedented,” says John Reed Stark, who spent 20 years as an attorney at the US Securities and Exchange Commission (SEC), including as the founder of its Office of Internet Enforcement. “If they’re a drug dealer or a terrorist or a child pornography peddler, they’re going to get caught.” He describes Binance’s agreement as a “24/7, 365-days-a-year financial colonoscopy.”

One US prosecutor, who asked not to be named because they weren’t authorized to speak to media about the case, calls the degree of access to Binance’s records described in the agreement “kind of crazy,” and remains in disbelief at the idea of Binance abiding by the settlement. “I don’t know what kind of business would want to operate while allowing that much government oversight, especially one that’s deliberately stayed out of the US so that they’re not under our nose,” they say. “The other option must have been really bad.”

The holidays are around the corner. These next few weeks might feel a little hectic, but sipping on a warm, cozy cup of tea is always the answer when you’re looking for a sense of peace amidst the madness. You’re in luck because we’ve found some great deals on our favorite electric kettles, plus WIRED-tested mugs and tumblers. Not a tea drinker? Don’t worry, most of these work well for coffee lovers too.

Special offer for Gear readers: Get WIRED for just $5 ($25 off). This includes unlimited access to WIRED.com, full Gear coverage, and subscriber-only newsletters. Subscriptions help fund the work we do every day.

If you buy something using links in our stories, we may earn a commission. This helps support our journalism. Learn more.

Electric Kettle Deals

Hay Sowden Electric Kettle

Photograph: Hay

Hay’s Sowden is more about aesthetics than function. Its casing feels cheap, the opening mechanism is clunky, and there’s no way to set a specific temperature. The boil speed is fairly average too. But there’s no denying it’s beautiful—this kettle is often spotted in magazine features and interior design photoshoots. If all you want is boiling water, it’ll do it just fine in style on any countertop.

The PerfecTemp comes with preset buttons for white tea, green tea, oolong tea, black tea, and a “delicate” button for light and spring teas that boils water to 160 degrees—so this is especially great for tea lovers. If you’re in the mood for coffee, there’s a preset for French press coffee as well. Other notable features include a backlit water window (to help you see how much water you’re filling it with), a blue light that indicates when it’s boiling, and a mesh filter near the pouring spout you can remove to clean. This discount only applies to the stainless steel version, but this is the lowest price we’ve seen yet.

KitchenAid 1.5 L Pro Line

Photograph: KitchenAid

We recommend this kettle for anyone looking for a quieter boil, as it’s noticeably fainter than the competition. It has a gentle alarm tone when it’s finished boiling too. There’s a keep-warm function, dual wall insulation (it remains cool to the touch while the water is hot), a sliding 50-to-100-degree-Celsius adjustable temperature gauge so you can set the temp, plus another dial on the kettle so you can see how hot the water is at a glance.

If you’re looking for a glass kettle, this one is easy to use, durable, and affordable (even without the discount). Made of borosilicate glass, it can withstand high temperatures and the rounded handle allows for a sturdy hold and a smooth pour. The built-in blue LED on the bottom of the kettle turns on when the water is boiling and off when it’s done (it also automatically shuts off). It’s on the louder side, but it boils water quickly—we boiled a little over one liter of water in only four minutes.

Mug and Tumbler Deals

East Fork The Mug

Photograph: East Fork

We love these mugs from East Fork. The ceramic feels strong and sturdy, and the gently sloped grip is large enough for most hands. It comes in a variety of unique colors, including Amaro, Morel, and Panna Cotta (certain colors do sell out quickly though). This discount applies to the four-pack.

The Carter Mug is too wide for most cars’ cupholders, but its width allows you to use a majority of manual coffee brewers (like the Hario pour-over or an AeroPress) directly on it. The interior is ceramic-coated, so you won’t get any metallic flavors, while the textured matte exterior coating is durable. It’s also fully leak-proof, so you can throw it into your bag even when it’s full. This discount applies to select colors including Warm Pink, Slate Grey, Corduroy Red, Buttered Popcorn, and Mint Chip. The smaller size is also on sale.

Sttoke Reusable Ceramic Travel Mug

Photograph: Stokke

If you want a more stylish travel mug, this one from Sttoke is an excellent choice. It’s made from a shatter-resistant ceramic, has a high-end matte finish and ergonomic grip, and comes complete with an easy-to-use slide lid. Sttoke also claims it can keep hot drinks hot for three hours and cold drinks cold for six hours.

Yeti’s Rambler pint has a matte exterior that’s enjoyable to hold and a clear lid that allows you to see how much liquid is in your cup (you can also swap it out for Yeti’s lid with a sliding magnetic closure or a straw lid). It’ll also keep your coffee or tea hot for several hours. It’s dishwasher-friendly too. They’re stackable for easy storage.

WIRED senior editor Michael Calore uses this vacuum flask to keep his freshly brewed tea hot all afternoon. According to Hydro Flask, the double-walled vacuum insulation can keep hot drinks hot for 12 hours. In testing, we found it was still boiling hot after four or five hours. You can use the Hydro Flask to brew fresh cups of tea individually too—simply fill it with piping hot water and pour out a cup at a time onto your tea leaves or teabag.

While Wyden’s letter says that governments outside the US have requested people’s push notification records, the Federal Bureau of Investigation (FBI) has done so as well. A February 2021 search warrant application submitted by an FBI agent to the US District Court in Washington, DC, requested details for two accounts controlled by Meta, (then Facebook), specifically citing a request for push notification tokens. The search warrant request related to an investigation into a person accused of taking part in the January 6, 2021, attack on the US Capitol.

Meta, which owns Facebook, WhatsApp, and Instagram, did not immediately respond to WIRED’s request to comment. The DOJ has not yet responded to a request for comment. A spokesperson for Signal, the popular encrypted messaging app, also did not respond.

Although Wyden is asking the DOJ to allow Apple and Google to discuss government requests for push notification records, the senator’s letter appears to have enabled them to do just that.

An Apple spokesperson tells WIRED that the company has updated its Law Enforcement Guidelines in its transparency report to reflect government requests for push notification records. The company will also begin to detail these requests in its next transparency report.

“Apple is committed to transparency and we have long been a supporter of efforts to ensure that providers are able to disclose as much information as possible to their users,” Apple says in a statement. “In this case, the federal government prohibited us from sharing any information and now that this method has become public we are updating our transparency reporting to detail these kinds of requests.”

Google confirmed to WIRED that it receives requests for push notification records, but the company says it already includes these types of requests in its transparency reports.

“We were the first major company to publish a public transparency report sharing the number and types of government requests for user data we receive, including the requests referred to by Senator Wyden,” a Google spokesperson tells WIRED. “We share the senator’s commitment to keeping users informed about these requests.”

A WIRED review of Google’s most recent transparency report for the period between December 2019 and December 2022 found that it does not specifically break out government requests for push notification records, and Google confirmed that it aggregates this data in its transparency report.

Google’s transparency report shows that the US government requested Google Cloud Platform data from enterprise customers 175 times during the period, and of those used a search warrant 13 times. It is unclear whether any of those requests for user data included push notification records—details that may, following Wyden’s letter, be revealed in the future.

Additional reporting by William Turton and Dhruv Mehrotra.

Google says there are three versions of Gemini: Ultra, the largest and most capable; Nano, which is significantly smaller and more efficient; and Pro, of medium size and middling capabilities.

From today, Google’s Bard, a chatbot similar to ChatGPT, will be powered by Gemini Pro, a change the company says will make it capable of more advanced reasoning and planning. Today, a specialized version of Gemini Pro is being folded into a new version of AlphaCode, a “research product” generative tool for coding from Google DeepMind. The most powerful version of Gemini, Ultra, will be put inside Bard and made available through a cloud API in 2024.

Sissy Hsiao, vice president at Google and general manager for Bard, says the model’s multimodal capabilities have given Bard new skills and made it better at tasks such as summarizing content, brainstorming, writing, and planning. “These are the biggest single quality improvements of Bard since we’ve launched,” Hsiao says.

New Vision

Google showed several demos illustrating Gemini’s ability to handle problems involving visual information. One saw the AI model respond to a video in which someone drew images, created simple puzzles, and asked for game ideas involving a map of the world. Two Google researchers also showed how Gemini can help with scientific research by answering questions about a research paper featuring graphs and equations.

Collins says that Gemini Pro, the model being rolled out this week, outscored the earlier model that initially powered ChatGPT, called GPT-3.5, on six out of eight commonly used benchmarks for testing the smarts of AI software.

Google says Gemini Ultra, the model that will debut next year, scores 90 percent, higher than any other model including GPT-4, on the Massive Multitask Language Understanding (MMLU) benchmark, developed by academic researchers to test language models on questions on topics including math, US history, and law.

“Gemini is state-of-the-art across a wide range of benchmarks—30 out of 32 of the widely used ones in the machine-learning research community,” Collins said. “And so we do see it setting frontiers across the board.”

Demis Hassabis has never been shy about proclaiming big leaps in artificial intelligence. Most notably, he became famous in 2016 after a bot called AlphaGo taught itself to play the complex and subtle board game Go with superhuman skill and ingenuity.

Today, Hassabis says his team at Google has made a bigger step forward—for him, the company, and hopefully the wider field of AI. Gemini, the AI model announced by Google today, he says, opens up an untrodden path in AI that could lead to major new breakthroughs.

“As a neuroscientist as well as a computer scientist, I’ve wanted for years to try and create a kind of new generation of AI models that are inspired by the way we interact and understand the world, through all our senses,” Hassabis told WIRED ahead of the announcement today. Gemini is “a big step towards that kind of model,” he says. Google describes Gemini as “multimodal” because it can process information in the form of text, audio, images, and video.

An initial version of Gemini will be available through Google’s chatbot Bard from today. The company says the most powerful version of the model, Gemini Ultra, will be released next year and outperforms GPT-4, the model behind ChatGPT, on several common benchmarks. Videos released by Google show Gemini solving tasks that involve complex reasoning, and also examples of the model combining information from text images, audio, and video.

“Until now, most models have sort of approximated multimodality by training separate modules and then stitching them together,” Hassabis says, in what appeared to be a veiled reference to OpenAI’s technology. “That’s OK for some tasks, but you can’t have this sort of deep complex reasoning in multimodal space.”

OpenAI launched an upgrade to ChatGPT in September that gave the chatbot the ability to take images and audio as input in addition to text. OpenAI has not disclosed technical details about how GPT-4 does this or the technical basis of its multimodal capabilities.

Playing Catchup

Google has developed and launched Gemini with striking speed compared to previous AI projects at the company, driven by recent concern about the threat that developments from OpenAI and others could pose to Google’s future.

At the end of 2022, Google was seen as the AI leader among large tech companies, with ranks of AI researchers making major contributions to the field. CEO Sundar Pichai had declared his strategy for the company as being “AI first,” and Google had successfully added AI to many of its products, from search to smartphones.