Meta’s new release sparks debate about open versus closed source AI: Hashtag Trending for Friday, April 19, 2024

Share post:

Just how real is quantum computing? We have an amazing guest on our Weekend Edition who will talk about how she is helping people prepare for IT careers using quantum computing.

Meta’s new AI release sparks a debate about open versus closed source AI, major legislation expanding US government surveillance capabilities goes largely unnoticed, big questions about how accurate these AI launch videos are, and before you book that next business trip, a former Boeing manager says he won’t fly on some Boeing planes.

All this and more on the “flying under the radar” edition of Hashtag Trending. I’m your host, Jim Love. Let’s get into it.


Meta has released a compact version of its latest open-source AI model called Llama 3, as well as an updated version of its AI assistant that can now answer questions using real-time web information.

These launches have reignited the debate around the future control and accessibility of powerful AI systems.

The company says the new Meta AI assistant, available across its apps like WhatsApp and Instagram as well as a website, is the “most intelligent” free AI aid of its kind. It includes new capabilities like generating custom images and animated GIFs based on text prompts.

But in a separate but related interview, Meta’s chief AI scientist Yann LeCun argued for these types of AI platforms and models to remain open-source and decentralized – warning of the dangers of having a small number of companies control everyone’s “digital diet.”

LeCun said “Eventually all our interactions with the digital world will be mediated by AI assistants…this means they will constitute a shared infrastructure like the internet. We cannot have a small number of AI assistants controlling what everybody sees – this will be extremely dangerous for diversity of thought, for democracy, for just about everything.”

LeCun advocated for a future with many different open AI models that can be customized for various cultures, languages and use cases – preventing what he called monopolistic “echo chambers” over AI-powered knowledge.

The push reflects a philosophical divide in the AI industry, with some firms like OpenAI keeping their models closed and proprietary, while others position themselves as open-source champions.

Meta stated goals include not just making their model open, but making its latest model more multilingual over time. LeCun also argued the importance of diverse AI assistants emerging to reflect society’s diversity.

As these systems become conduits for how people interact with digital information and services, the debate around centralized control versus democratized access of this transformational technology will certainly intensify.

Sources include: Axios and Analytics India

For anyone following US political news you may have missed this story in the light of the trials, impeachment inquiries and the crisis over Ukraine funding, but there is a major debate and the U.S. Senate is set to vote on Thursday on renewing and potentially expanding controversial government surveillance powers.

Privacy advocates are sounding alarms over proposed changes that could compel a vast range of companies and individuals to hand over electronic communications to intelligence agencies.

At issue is the reauthorization of Section 702 of the Foreign Intelligence Surveillance Act, or FISA. This provision currently allows U.S. spy agencies to conduct warrantless surveillance of communications involving foreigners outside America who may pose national security threats.

FISA has long been criticized for also sweeping up data on American citizens, which intelligence agencies can then search domestically without a warrant – a practice that civil liberties groups consider unconstitutional.

Last week, the House passed a bill reauthorizing Section 702 for 8 more years. But it contained an amendment drastically broadening who could be legally defined as an “electronic communications service provider” – a classification that can compel companies to share private user data.

Oregon Senator Ron Wyden warned this could conscript millions into becoming “agents for Big Brother” – from office cleaners to any business with computer servers or wifi routers on their premises.

Wyden noted “The legislation gives the government unchecked authority to order millions of Americans to spy on behalf of the government…anyone with access to a server, wire, cable box, router, phone or computer.”

Privacy advocates like the American Civil Liberties Union and tech industry groups are urging the Senate to reject this provision, citing risks to digital privacy and US competitiveness if user data is exposed by government demands.

The White House and others maintain that in a world full of terrorist threats and instability that FISA provides tools necessary to intercept threats and protect American citizens.

As the Senate deliberates the renewal, it will need to weigh enhanced national security powers against public unease over eroding civil liberties protections in the digital age.

Given the other hub-bub and the speed at which this is moving, it’s unlikely that anyone will get a chance to hear a clear and reasoned debate of these exceptionally important issues.

Sources include: The Register

And for those whose job or even personal lives take them into airplanes for travel, we’ve covered apps that tell you which type of plane you might be one.

For those who think that’s just alarmist, you might want to check out another US senate subcommittee where a former Boeing manager, Ed Pierson has repeatedly talked about safety issues on the Boeing 737 Max jet like the one involved in the recent Alaska Airlines incident where a door blew off in midair.  He said that, once, when he realized he had been booked on a 737 Max, he got off before the plan could take off.

When the 737 Max 9 lost its door in midair, leaving a gaping hole in the plane, the National Transportation Safety board found that bolts designed to secure it were missing. The same report said that they door plug had been removed in a Boeing factory to fix some broken rivets but Boeing told the board that it didn’t have documentation for this work.

Pierson testified that a whistle-blower at Boeing gave him documents that indicate a “criminal cover-up” related to the door incident. Boeing has reportedly claimed that there’s no documents of work done on the door plug that came off the 737 Max jet.

Pierson said on Wednesday: “Records do in fact exist. I know this because I personally passed them to the FBI.”

Pierson was a senior manager at Boeing’s 737 factory and retired in 2018 before the first Boeing 737 Max 8 crash.

Pierson said, “I’m not gonna sugarcoat this, this is a criminal conspiracy.”

The FBI is looking into whether criminal charges should be brought in this case and passengers from the Alaska Airlines flight were reportedly sent letters from the FBI saying they might be victims of a crime.

For anyone who has ever sat beside an exit door, it adds a totally different meaning to that speech the flight crew makes when they ask if you are “prepared to act in the event of an emergency.”

Sources include: Business Insider

Major tech companies like Amazon are facing scrutiny over claims that some of their highly touted artificial intelligence systems are actually relying heavily on offshored human labor. Critics argue this amounts to traditional outsourcing being repackaged under the banner of AI. Meanwhile, the companies deny the allegations, saying their AI capabilities are indeed automated and that human reviewers play only a limited role.

Controversy erupted recently when reports emerged that Amazon’s “Just Walk Out” cashier-less checkout system utilizes human workers in India reviewing surveillance footage to verify purchases. This raised accusations that the tech giant had overstated the AI automation behind this system that was marketed as allowing customers to simply grab items and leave the store.

In a recent article, author Janet Vertesi argues that much of what is being branded as AI is really just old-fashioned labor outsourcing in a new guise.

Vertesi said, “AI is just today’s buzzword for outsourcing, and it comes with the same problems that have plagued outsourced companies for decades…behind the curtain is the familiar phenomenon of outsourcing – expensive skilled labor traded for cheap, unskilled labor abroad.”

However, Amazon has forcefully pushed back on this perception. The company’s VP overseeing Just Walk Out, Jon Jenkins, told Axios in a recent interview that human reviewers in India only analyze a “small percentage” of cases after the fact to improve the AI’s accuracy – not watch live shoppers.

Jenkins said, “This notion that there are human reviewers watching live shoppers – that is completely not true…way less than 1,000 people help make sure automatically generated receipts are accurate.”

Jenkins argues Just Walk Out utilizes advanced sensors, cameras and AI to enable the grab-and-go experience, while admitting there’s still work to do in scaling the technology efficiently across more locations.

The debate speaks to the broader challenge of separating AI hype from reality. A number of AI products, not just Amazon, have let’s say “enhanced the performance of their systems” by editing the video. Google got caught editing one of its AI launches to enhance the appeared performance. Yesterday, we covered scathing reviews of a of Humane’s AI pendant where the actual performance of the device was nothing like what the video demo showed – and this infuriated reviewers. When you start digging, there are many examples to be found.

But the Amazon case also raised concerns around tech firms cutting costs by outsourcing labor overseas under the veneer of automation.

As AI enables even more sophisticated simulations, look for this problem to intensify. In answer to what Groucho Marx said, “are you going to believe me or your lying eyes?” Maybe there is another saying, “there ought to be a law…”

Sources include: Techpolicy.press and Axios

And that’s our show for today…

Hashtag trending goes to air five days a week with a weekend interview show. And we are also on YouTube.

Find us at our new home at technewsday.ca or .com – you pick. And you can reach me with comments, suggestions or even criticism at therealjimlove@gmail.com or at editorial@technewsday.ca

I’m your host Jim Love, have a Fabulous Friday.

 

 

 

 

 

 

 

The post Meta’s new release sparks debate about open versus closed source AI: Hashtag Trending for Friday, April 19, 2024 first appeared on IT World Canada.

SUBSCRIBE NOW

Related articles

Pablo Listingart from ComIT – Hashtag Trending the Weekend Edition for May 25, 2024

Pablo Listingart, founder and executive director of ComIT, discusses the resource shortage in cybersecurity and IT and the...

Cyber Security Today, May 24, 2024 – A threat actor leverages Windows BitLocker in ransomware attacks, beware of ORB networks, and more

A threat actor leverages Windows BitLocker in ransomware attacks, beware of ORB networks, and more. Welcome to Cyber Security...

Consumers don’t trust social media companies with AI. Hashtag Trending for Friday, May 24th, 2024

A new Axios poll says that consumers don’t trust social media giants branching into AI.  Two companies headed...

Cyber Security Today, May 22, 2024 – LockBit ransomware gang hits more victims, Fluent Bit servers need to be updated, and more

LockBit ransomware gang hits more victims, Fluent Bit servers need to be updated, and more. Welcome to Cyber Security...

Become a member

New, Relevant Tech Stories. Our article selection is done by industry professionals. Our writers summarize them to give you the key takeaways