Meta releases beefed-up AI models, eyes integration into its apps

By weaving AI into its family of apps, Meta will quickly get features powered by the technology to billions of people and benefit from seeing what users do with it. (AFP/File)
Short Url
Updated 19 April 2024
Follow

Meta releases beefed-up AI models, eyes integration into its apps

  • AI model Llama 3 takes step towards human-level intelligence, Meta claims
  • Company also announced new AI Assistant integration into its major social media apps

SAN FRANCISCO: Meta on Thursday introduced an improved AI assistant built on new versions of its open-source Llama large language model.
Meta AI is smarter and faster due to advances in the publicly available Llama 3, the tech titan said in a blog post.
“The bottom line is we believe Meta AI is now the most intelligent AI assistant that you can freely use,” Meta co-founder and chief executive Mark Zuckerberg said in a video on Instagram.
Being open source means that developers outside of Meta are free to customize Llama 3 as they wish and the company may then incorporate those improvements and insights in an updated version.
“We’re excited about the potential that generative AI technology can have for people who use Meta products and for the broader ecosystem,” Meta said.
“We also want to make sure we’re developing and releasing this technology in a way that anticipates and works to reduce risk.”
That effort includes incorporating protections in the way Meta designs and releases Llama models and being cautious when it adds generative AI features to Facebook, Instagram, WhatsApp, and Messenger, according to Meta.
“We’re also making Meta AI much easier to use across our apps. We built it into the search box right at the top of WhatsApp, Facebook, and Instagram messenger, so any time you have a question, you can just ask it right there,” said Zuckerberg in the video.
AI models, Meta’s included, have been known to occasionally go off the rails, giving inaccurate or bizarre responses in episodes referred to as “hallucinations.”
Examples shared on social media included Meta AI claiming to have a child in the New York City school system during an online forum conversation.

Meta AI has been consistently updated and improved since its initial release last year, according to the company.
“Meta’s slower approach to building its AI has put the company behind in terms of consumer awareness and usage, but it still has time to catch up,” said Sonata Insights chief analyst Debra Aho Williamson.
“Its social media apps represent a massive user base that it can use to test AI experiences.”
By weaving AI into its family of apps, Meta will quickly get features powered by the technology to billions of people and benefit from seeing what users do with it.
Meta cited the example of refining the way its AI answers prompts regarding political or social issues to summarize relevant points about the topic instead of offering a single point of view.
Llama 3 has been tuned to better discern whether prompts are innocuous or out-of-bounds, according to Meta.
“Large language models tend to overgeneralize, and we don’t intend for it to refuse to answer prompts like ‘How do I kill a computer program?’ even though we don’t want it to respond to prompts like ‘How do I kill my neighbor?’,” Meta explained.
Meta said it lets users know when they are interacting with AI on its platform and puts visible markers on photorealistic images that were in fact generated by AI.
Beginning in May, Meta will start labeling video, audio, and images “Made with AI” when it detects or is told content is generated by the technology.
Llama 3, for now, is based in English but in the coming months Meta will release more capable models able to converse in multiple languages, the company said.


Russian rights group says YouTube threatens to block its anti-war channel

Updated 23 May 2024
Follow

Russian rights group says YouTube threatens to block its anti-war channel

  • OVD-Info said YouTube’s email followed a request by the Russian communications regulator Roskomnadzor

LONDON: A leading Russian rights group says it has received a notice from YouTube threatening to block access in Russia to one of its video channels featuring news on the war in Ukraine.
OVD-Info, an independent protest monitoring network, said it had received an email from YouTube in early May saying that the Russian communications regulator Roskomnadzor had found content on the channel that violated a law on information technology.
“If you do not remove the content, Google may be required to block it,” Alphabet Inc’s YouTube wrote, according to screenshots of an email shared with Reuters. The email did not specify which part of the law OVD-Info was accused of violating.
The channel, Kak Teper (What’s Going On), has 100,000 subscribers and features interviews with Russian opposition figures and political news segments that often touch on the war.
“We are consulting with YouTube and Google and trying to explain that the demand to block our channel is an act of political censorship,” said OVD-Info spokesperson Dmitrii Anisimov. He said the group’s other YouTube channel was not affected.
Contacted by Reuters two times about Youtube’s discussions with OVD-Info, a spokesperson for YouTube did not respond. The spokesperson answered separate questions about three other opposition channels which had videos blocked.
Like many Western technology companies, Google quit Russia over its invasion of Ukraine, pulling its staff and suspending all advertising sales, including on YouTube.
Russia has blocked the vast majority of foreign social media platforms, but has stopped short of blocking YouTube despite fining the video-hosting platform repeatedly for failing to delete content Moscow deems illegal.
YouTube has tens of millions of monthly users in Russia and blocking the entire platform could prove highly unpopular.
Russian independent media reported on Monday that YouTube had deleted videos from three other channels that provided information on how to evade Russian military service.
Two of the groups told Reuters their content had been reinstated within a day after the media reports.
When contacted by Reuters about the videos, the YouTube spokesperson replied by email: “The content in question has been reinstated to YouTube,” without elaborating.
OVD-Info’s Kak Teper would be the first Russian human rights channel to be banned on YouTube, as opposed to just a few videos, according to Natalia Krapiva, tech legal counsel at global digital rights non-profit Access Now.
“We will not have any YouTube to fight for anymore if all the civil society is blocked there,” Krapiva said in a phone interview.


Dutch prosecutors studying complaint against Booking.com’s Israeli settlement listings

Updated 23 May 2024
Follow

Dutch prosecutors studying complaint against Booking.com’s Israeli settlement listings

AMSTERDAM: Dutch prosecutors are looking into a criminal complaint against Booking.com over its listing of rental properties in Israeli settlements, they said on Thursday.
Dutch non-profit organization SOMO said it had filed the complaint with the Dutch public prosecutor in November, together with three other human rights groups, but had not gone public with it before.
In their complaint the groups accuse Booking.com of “profiting from war crimes by facilitating the rental of vacation homes on land stolen from the indigenous Palestinian population.”
Prosecutors were studying the complaint, but could not give a timeline for a decision on possible further steps, spokesperson Brechje van de Moosdijk said.
Booking in a response said it disagreed with the allegations and that there are no laws prohibiting listings in Israeli settlements, while a range of US state laws would prohibit divesting from the region.
“Legal action has been taken against other companies that have tried to withdraw their activities, and we would expect the same to happen in our case,” a spokesperson for the company said.
SOMO said its research had shown that Booking’s platform offered up to 70 listings for properties in East Jerusalem and the occupied West Bank between 2021 and 2023.
It argued that revenues acquired from renting out those properties are “proceeds of criminal activities,” and that by booking these proceeds in the Netherlands the company is violating Dutch anti-money laundering rules.
The settlements built on land captured by Israel in the 1967 Middle East war are deemed illegal by most countries, including the Netherlands. Their presence is one of the fundamental issues in the Israeli-Palestinian conflict.
Palestinians seek to establish an independent state in the West Bank and Gaza Strip with East Jerusalem as their capital. Israeli settlers cite Jewish historic connections to the land.


News Corp. makes deal to let OpenAI use its content

Updated 23 May 2024
Follow

News Corp. makes deal to let OpenAI use its content

  • ChatGPT’s creator is also in the process of signing content licensing agreements with media outlets

NEW YORK: News Corp. on Wednesday announced a deal to let ChatGPT-maker OpenAI use content from its publications in artificial intelligence products.
OpenAI will get access to current and archived content from News Corp. properties including The Wall Street Journal, Barron’s, MarketWatch, and The New York Post, according to a joint release.
Financial terms of the deal were not disclosed, but the Wall Street Journal cited sources close to the company as saying it was valued at more than $250 million over five years and included credits for News Corp. using OpenAI technology.
Authors, artists, and news groups have been accusing OpenAI and its rivals in the generative artificial intelligence market of using copyrighted content for training models without asking permission or paying.
Generative AI models are trained on mountains of data in the effort to get software to think the way people do.
“This landmark accord is not an end, but the beginning of a beautiful friendship in which we are jointly committed to creating and delivering insight and integrity instantaneously,” News Corp. chief executive Robert Thomson said.
OpenAI gets permission to display News Corp. content in response to queries by users of its technology, according to terms of the deal.
“Our partnership with News Corp. is a proud moment for journalism and technology,” Open AI CEO Sam Altman said in the release.
“Together, we are setting the foundation for a future where AI deeply respects, enhances, and upholds the standards of world-class journalism.”
ChatGPT’s creator is also in the process of signing content licensing agreements with media outlets — including the Associated Press, Germany’s Axel Springer Group (publisher of tabloid Bild), French daily Le Monde and Spanish conglomerate Prisa Media — to enrich its models.
The announcement of the agreement with News Corp. comes on the heels of a new controversy, after actress Scarlett Johansson accused OpenAI of copying her voice for a new voice assistant without her permission.
Altman has apologized and announced the suspension of the voice, called “Sky.”


Antisemitism group posts fake news about politician after Ireland recognizes Palestinian state

Updated 23 May 2024
Follow

Antisemitism group posts fake news about politician after Ireland recognizes Palestinian state

  • Stop Antisemitism puts message on X claiming Irish Foreign Minister Micheal Martin’s daughter was kidnapped and raped in Gaza on Oct. 7; later admits this ‘did not actually occur’
  • Blatant disinformation outrages users; some suggest such posts only provoke antisemitism, others say comments about rape should not be made lightly

DUBAI: US-based organization Stop Antisemitism posted a message on social media platform X on Wednesday that appeared to state Irish Foreign Minister Micheal Martin’s daughter, Aoibhe Martin, had been kidnapped and raped by Hamas in Gaza on Oct. 7 and that now “he is rewarding his daughter’s rapists with a state of their own.”

The organization added another post, more than an hour later, in which it said the initial post “is for illustrative purposes only” and the events it described “did not actually occur.”

 

 

The blatant use of disinformation outraged many X users, with some suggesting that such posts serve only to increase incidents of antisemitism. Others said comments about rape should not be made lightly and that there was nothing “illustrative” about the post.

 

Critics say that disinformation and fake news has greatly increased on X since Elon Musk bought the platform in April 2022. In the past two years, the company has shed thousands of jobs, many of them related to content moderation.

European Commission Vice President Vera Jourova last year accused X of being the social media platform with the highest ratio of fake news, and urged Musk to comply with EU laws designed to combat disinformation.

In April, X’s own artificial intelligence chatbot Grok generated a fake headline that stated: “Iran Strikes Tel Aviv with Heavy Missiles.” It was promoted on the main X feed.

In the 48 hours following the Oct. 7 attacks by Hamas on Israel, misinformation was rampant on the platform. One video that claimed to show Israeli generals captured by a Hamas fighter was actually footage of separatists detained in Azerbaijan. Another clip showing an airplane being shot down was accompanied by the hashtag #PalestineUnderAttack when it was really footage taken from the video game Arma 3. The former video was viewed more than 1.7 million times in two days, the latter more than 500,000 times.

Earlier on Wednesday, Martin had announced in a video message posted on X that the Irish government will formally recognize the State of Palestine on May 28.

“The Palestinian people’s right to self-determination, peace, dignity and statehood must be vindicated,” he added. “It is our conviction that the two-state solution remains the only viable option to secure a just and lasting peace that fulfills these rights for both Israelis and Palestinians alike.”

He added that recognition of Palestine as a state did not mean the legitimization of Hamas.

“Recognition does not involve recognition of a government, it’s recognition of a state,” he told Irish radio program The Pat Kenny Show.

Martin had not responded to Stop Antisemitism’s post on X at the time of writing.


Spotify spotlights Khaleeji music in New York’s Times Square

Updated 23 May 2024
Follow

Spotify spotlights Khaleeji music in New York’s Times Square

  • Saudi artist Sultan Al-Murshed and Iraqi artist Aseel Hameem have been selected as Spotify’s RADAR Arabia and EQUAL Arabia artists for May

DUBAI: Spotify is spotlighting Saudi artist Sultan Al-Murshed and Iraqi artist Aseel Hameem, who have been selected as Spotify’s RADAR Arabia and EQUAL Arabia artists for May, in New York’s Times Square.

“We continue to be committed to showcasing and celebrating genres and creators reflecting different spectrums of Arabic music,” said Nada Elmeri, Spotify’s Artist and Label Partnerships Manager for the Gulf Region at Spotify MENA.

This month was dedicated to celebrating Khaleeji Pop — a genre “that has played a pivotal role in our childhood memories yet continues to resonate with young listeners and is met with a lot of loyalty from fans across different generations,” Elmeri told Arab News.

Al-Murshed, a rising star from Saudi Arabia, was selected as this month’s RADAR Arabia artist for winning listeners over with his melodies and vocals.

His debut single “Wala Ghaltah,” released in 2022, has amassed over 1 million streams on Spotify. Over the last month or so, his streams have increased by 73 percent and fans have saved his music 97 percent more over the same period.

He also worked with renowned DJ and producer R3HAB and Big Bo in 2022 for the official Gamers8 anthem, “Challenge.”

This month’s EQUAL Arabia artist is Aseel Hameem, daughter of renowned Iraqi musician Kareem Hameem, who began her musical journey when she was a young girl. Her talent was evident quickly garnering her the nickname “The Guitar of Iraq.”

Despite her Iraqi roots, Hameem has mastered singing in the Saudi and Khaleeji dialects, gaining substantial support in Saudi Arabia. Her most popular hits include “Shkad Helw” and “Al Mafrod” with the latter garnering over 14 million streams on Spotify.

Her latest release, “Mostafz Alnas,” has resonated with audiences in Saudi Arabia, Egypt, Iraq, Germany, and the US, Spotify said.

Throughout May, Spotify is running a promotional campaign to boost Hameem’s work including editorial placements and social media support.

Both Al-Murshed and Hameem were featured on a billboard in New York City’s Times Square as well as on the covers of the RADAR Arabia and EQUAL Arabia playlists on Spotify this month.

“To spotlight the (Khaleeji Pop) genre in an impactful way, we featured two Khaleeji artists across different career journeys,” said Elmeri.

Al-Murshed “represents the new wave of the genre” while Hameem “has been a force over the years with her presence visible on our Saudi Wrapped lists,” she added.

RADAR Arabia and EQUAL Arabia are Spotify’s global music programs aimed at supporting emerging artists and female artists respectively.