Skip to main content
All Stories Tagged:

Artificial Intelligence

Artificial intelligence is more a part of our lives than ever before. While some might call it hype and compare it to NFTs or 3D TVs, AI is causing a sea change in nearly every facet of life that technology touches. Bing wants to know you intimately, Bard wants to reduce websites to easy-to-read cards, and ChatGPT has infiltrated nearly every part of our lives. At The Verge, we’re exploring all the good AI is enabling and all the bad it’s bringing along.

Microsoft says its automated AI red teaming tool finds malicious content “in a matter of hours.”

PyRIT, or Python Risk Identification Toolkit, can point human evaluators to “hot spot” categories in AI that might generate harmful prompt results.

Microsoft used PyRIT while redteaming (the process of intentionally trying to get AI systems to go against safety protocols) its Copilot services to write thousands of malicious prompts and score the response based on potential harm in categories that security teams can now focus on.

A New Orleans magician says he made the AI Biden robocall telling people not to vote.

Paul Carpenter told NBC News that he was paid $150 to produce a fake AI-generated voice message from Joe Biden — and that the political operative Steve Kramer hired him to do it. Kramer has worked with Democratic presidential candidate Dean Phillips’ campaign to gather signatures to get on the ballot. NBC didn’t find evidence that the Phillips campaign was involved in the robocalls.

The parties reportedly involved apparently were not very discreet. Kramer texted the magician, “Have AI voice project” in September. And then there’s this:

On Jan. 22, when NBC News first broke the news of the fake Biden robocall, Steve Kramer texted Carpenter a link to the story along with the message, “Shhhhhhh,” to which Carpenter replied, “Gtfooh,” an acronym used to express astonishment. 

External Link
Staring carefully into surveillance camera: “Would you... like... fries with that?”

AI is theoretically a labor-saving tool, but here’s yet another example of a perhaps more immediate use: keeping tabs on workers, in this case so they’ll upsell more customers at fast-food joints. It’s even got a cute name:

Riley, installed at about 100 stores across the country including KFC and Taco Bell franchisees, ingests data including workers’ conversations with customers, and uses AI to detect whether and how often employees have tried to “upsell” (offered add ons or extra toppings), “upsize” (offered a larger size) or tried to sign a customer up for a loyalty program. Employees who make the most suggestions and whose suggestions get converted into sales receive cash bonuses based on a scorecard generated by Hoptix’s AI system.

External Link
Stability AI changes it up for Stable Diffusion 3.

VentureBeat reports that the next generation of Stability AI’s flagship AI image generation model will use a diffusion transformer framework similar to OpenAI’s Sora. Its current models rely on diffusion architecture alone.

The company said Stable Diffusion 3 — currently in previews — should be better at spelling (look closely at text in AI-generated images, and you know they don’t look right) and boost image quality.

External Link
Google is reportedly Reddit’s $60 million per year AI content licensing customer.

Charging for data access was a flash point for last year’s protests, with Reddit CEO Steve Huffman telling The Verge, “ licensing is a new potential business for us.”

Now, with Reddit’s IPO launch close at hand, Reuters is putting a name on rumors of an AI company that’s paying Reddit for training data, in a deal that could be a model for similar arrangements, citing three sources who say the company is Google.

From ChatGPT to Gemini: how AI is rewriting the internet

How we use the internet is changing fast thanks to the advancement of AI-powered chatbots that can find information and redeliver it as a simple conversation.

Gemini, Gemma, Goose.

These are just a few of the names that Google has bestowed upon its AI products in recent weeks.

Is it confusing for those of us on the outside to track all of these name changes? Yes. Yes, it is. And as Business Insider details in this fun story, even Google’s own employees are roasting the company by posting memes about it internally.

One such meme Googlers recently shared.
One such meme Googlers recently shared.
Business Insider
External Link
“Generative AI has hit the tipping point.”

As Nvidia reports its Q4 2023 earnings, CEO Jensen Huang says:

Accelerated computing and generative AI have hit the tipping point. Demand is surging worldwide across companies, industries and nations

That demand, and Nvidia’s dominance in AI chips, powered the company’s record earnings of $60.9 billion for the full year 2023 — a 126 percent increase from last year. Its Q4 revenue of $22.1 billion is an increase of an astounding 265 percent over the same period last year.

External Link
The Gray Lady is working on a generative AI tool for advertisers.

The New York Times ad team is testing an internal generative AI chatbot that will recommend the best placement for ad campaigns and target niche audiences, reports Axios.

For example, a car company looking to position itself as family-friendly might use the tool to align some of its messages with articles about technology and others to pet owners across the Times’ site and app.

Axios says the paper is “broadly” investigating AI use cases, despite suing OpenAI and Microsoft for copyright infringement, and the Times recently announced it is building a team to explore the use of generative AI on the editorial side.

External Link
Nvidia lets Google’s Gemma AI model loose on its GPUs.

The open-source Gemma models are optimized for “the installed base of over 100 million Nvidia RTX GPUs,” installed in PCs around the world, in addition to Nvidia’s ubiquitous AI chips like the H100.

The models will also be part of Nvidia’s Chat with RTX demo, which lets AI models run locally and access users’ files to generate answers to prompts.

External Link
Calls for regulating AI deepfakes are growing.

An open letter signed by AI researchers, including Algorithmic Justice League founder Joy Boulamwini and former presidential candidate Andrew Yang, said governments should look to fully criminalize deepfake child pornography even with fictional children, create criminal penalities for people who make and share “harmful” deepfakes, and require developers to be held liable if their safety measures are easily bypassed.

US policymakers have discussed regulating deepfakes, though mostly in the context of the upcoming elections. It’s rare for open letters to influence regulation, but AI is a fraught issue that some lawmakers might take these suggestions into account.

External Link
Gemini is now officially part of Google Workspace, too.

The best thing about Google’s Gemini model is the way it works with your Google stuff. So this is no surprise: Gemini is being built right into Workspace. The Gemini Enterprise plan costs $30 a month, and Gemini Business, with slightly fewer generative features, is $20 a month. (You can always pony up $20 a month for the Google One package with Gemini, too.) Bye-bye Duet AI, it’s just Gemini all the way down.

One month with Microsoft’s AI vision of the future: Copilot Pro

Copilot Pro is a $20 per month subscription that includes AI features in Office apps and better image generation tools.

External Link
Microsoft is developing server tech to make AI chips run faster.

The Information reports Microsoft’s server network cards, which are currently in development, could replace the ones offered by Nvidia and save the company cash. Server cards move data quickly between servers. These new server cards should not only improve the performance of Nvidia chips, but Microsoft’s upcoming Maia AI chips as well.

Sora can create video collages, too.

One of OpenAI’s employees showed off another of the company’s new text-to-video generator’s abilities.

This is some impressive AI creation of course, but what in blue blazes is happening in the upper right frame here?

External Link
SoftBank founder Masayoshi Son wants $100 billion for a new AI chip venture.

It’s code-named Izanagi and is designed to combat Nvidia, sources are telling Bloomberg. OpenAI’s Sam Altman had been courting Softbank and other companies to potentially bankroll a chipmaking arm already — but the Izanagi project is reportedly not connected to that.

If successful, Izanagi may become the largest investment in AI since Microsoft put a rumored $10 billion wager on OpenAI.

External Link
Microsoft, Meta, Google, and OpenAI agree to develop technology against election deepfakes.

As well as Amazon, Adobe, IBM, Arm, Stability AI, TikTok, X, Snap, LinkedIn, and seven other AI and tech companies in the AI Elections Accord.

By signing the accord, the companies promise to develop ways to catch deepfakes, asses models that could lead to deepfakes, catch and stop election deepfakes from spreading on social platforms, and be transparent with the public about false election information.

External Link
A California bill proposes creating a statewide AI oversight office.

State senator Scott Weiner proposes a Frontier Model Division in California’s Department of Technology that will oversee mandatory testing and review safety compliance of AI models. Weiner previously filed an “intent” bill, which was intended to kickstart discussions on regulating AI models in the state.

State Scoop reports that the bill does not clarify how it and the Frontier Model Division fit into California’s ongoing policy discussions on AI.