The AI Marathon - Pichai's Vision of U.S.-China Synergy in Technology

When you look at the landscape of artificial intelligence, it's clear: China and its AI engineers are stepping up. They're not just playing the game; they're redefining it. Sundar Pichai of Alphabet Inc. sees this. He knows that the future of AI isn't a solo journey. It's a collaboration, a dance of sorts, between giants like the U.S. and China.

At the Asia-Pacific Economic Cooperation conference in San Francisco, Pichai couldn't hide his awe at China's AI strides.

Now, think about this: it's not just tech giants at the table. We're talking about a convergence of minds from Microsoft, Citigroup, Tesla. And they're not just meeting with tech leaders. They're engaging with Presidents Xi Jinping and Joe Biden. This is more than a meeting; it's a melting pot of ideas, a forge where the future of AI is being shaped.

Alphabet CEO Sundar Pichai discusses a global consensus for AI regulation.

He speaks to @emilychangtv at the APEC CEO Summit https://t.co/WyRcSeuoae pic.twitter.com/jhxVSsxJmy

— Bloomberg TV (@BloombergTV) November 16, 2023

Pichai draws an interesting parallel here. AI and climate change – they're both global challenges that don't respect borders. A glitch in AI in one corner of the world can ripple across the globe. We're all in this interconnected web.

When asked about the principles of ethical behavior in an AI-driven world, Sundar Pichai emphasized, "I think it has to be grounded in the fundamental values of humanity, human rights, and universal human values."

And then there's Gemini, Google's new brainchild. This isn't just another AI. It's a leap into understanding and responding to us in our language. Google isn't just launching a product; they're birthing a new era. And the excitement? It's palpable. This is what Google was made for – moments like these, where the future isn't just imagined; it's built.

Photo: APEC CEO Summit USA 2023 / X

Read next: Galileo Claims GPT-4 Excels in Multitasking with Minimal Hallucinations Among LLMs
by Irfan Ahmad via Digital Information World

Galileo Claims GPT-4 Excels in Multitasking with Minimal Hallucinations Among LLMs

Galileo is a San Francisco based company that helps other companies build fine Large Language Model (LLM) apps. Galileo made a hallucination index that researches about different LLMs and shows which of them hallucinate when faced with multiple tasks at once. Hallucinating is a recurring problem observed in LLMs that result in inaccurate and baseless results. After the research conducted by Galileo, it was found that ChatGPT-4 hallucinates the least when it has to work on multiple tasks.
Galileo worked on different LLMs including Meta’s Llama, but it was soon revealed that many well known LLMs also cannot perform their tasks accurately without hallucinating. This keeps many enterprises from developing LLMs because they cannot work efficiently, especially in healthcare. Open AI’s GPT-4 was the only one that performed well when given tasks.

Identifying LLM Hallucination– How Galileo Checked the Performance of LLMs?

Many enterprises use AI and LLMs for their businesses. But when it comes to production, LLMs aren’t always reliable because sometimes their results aren’t 100%. It is because LLMs work according to the data they are given, and not if the information is factual or accurate. When generative AIs are being developed, there are a lot of things that need to be considered like is it for general use or is it a ChatBot for enterprises?

Companies use benchmarks to monitor LLMs performance, but there is no saying when they will hallucinate. To tackle this problem, co-founder of Galileo, Atindriyo Sanyal, chose 11 LLMs of all sizes and capacities and checked the performance of each of them one by one. He asked those LLMs general questions using TruthfulQA and TriviaQA to monitor their hallucination.

Galileo used its Correctness and Context Adherence metrics because they make it easy for engineers and data scientists to easily find out at what point LLMs are hallucinating. The metrics were focused on logic-based mistakes made by those LLMs.

Correctness Score of LLMs by Galileo

After QnA with LLMs, Galileo came to a conclusion about those LLMs with a correctness score. The LLM with the highest score was GPT-4 (0613) with a 0.77 score. GPT-3.5 Turbo(1106) was just behind with 0.74 score. GPT-3.5 Turbo Instruct and GPT-3.5 Turbo(0613) both had 0.70 correctness score. After the GPT versions came Meta’s Llama-2-70b with 0.65 score. Mosaic’s ML’s MPT-7b Instruct had a 0.40 score.

For information retrieval, GPT-4(0613) still came at the top with a 0.76 score. Zephyr-7b, a model by Hugging Face, surpassed LLama in this category with 0.71 score. Llama had 0.69 score.

The LLM that needed the most improvement and had the lowest points of 0.58 was Mosaic ML’s MPT-7b. UAE’s Falcon 40-b lies just above it with 0.60.

GPT-4(0163) was at the top for correctly generating long forms of texts like essays, articles etc. It has a score of 0.83. Llama came in second with 0.82 correctness score. These two models were the least to hallucinate when given multiple tasks. MPT-7b was the last with 0.53 score.

With Correctness Comes a Cost

As much as reliable ChatGPT-4 seems to be, there is still a factor that makes it less appealing to many–its pricing plan. GPT-4(0613) isn’t free to users but GPT 3.5 also serves almost the same as GPT-4 when it comes to being least hallucinated. There are also other alternatives like LLama-2-70b which is also good in its performance. Plus, it is free of cost.

Read next: AI's Data Dilemma Shows Less Can Be More
by Arooj Ahmed via Digital Information World

Choose Gifts Wisely: Mozilla Study Reveals Lack of Privacy Policies in Key Tech Products

Mozilla carried out research that reveals that over 100 tech gadgets shouldn’t be gifted to friends because of their privacy issues.

The good pals over at Mozilla published a guide named “Privacy Not Included” that gives a complete rundown of tech products that do not care for the user’s privacy. A lot of big tech companies' names were in the guide including Google, Microsoft and Amazon. For instance, the Federal Trade Commission fined Amazon $25 million because Alexa devices collected the data of many children. Similarly, Microsoft was fined with $20 because it was also collecting children’s data unlawfully.

Google also wasn’t left behind. The Fitbit Activity Tracker is used for tracking your heart rate, steps taken, skin temperature etc. Google has introduced Fitbit Charge 6 which will require your Google account for log in. This means all your Fitbit data will go to Google. By signing in to the Fitbit account, you won’t be accepting Fitbit’s privacy policy. You will have to accept the privacy policy of Google.

Sonos, which was everyone’s favorite, is also included in this dirty business. Mozilla found that it sells some of the users’ personal data for advertising purposes. This means it is ditching all of its privacy policies, without caring about the effects it will cause to their users. Bose is also stealing users’ information and giving it to third parties.

A kids’ smartwatch known as Angel Watch has no privacy policy. The Mozilla team tried hard to find any signs of privacy policy for the watch but it was nonexistent. The only privacy policy about Angel Watch is on its website. It states that when a person buys the watch or views the website, their personal data can be collected. This makes Angel watch suspicious because it makes youngsters vulnerable.

One of the most scary devices with no privacy Mozilla found out was Wyze Security Cameras. It was revealed that a glitch in the system left a lot of private feeds accessible to random people online. Tile Trackers, which are used to track things like pets, bags, etc via bluetooth, have added an Anti-Theft Mode that hides bluetooth devices. This makes it accessible to stalkers.

Apart from all the tech devices that were included in the Privacy Not Included list, Mozilla also appreciated some devices which respect their users’ privacies. One of them in Garmin which updated their privacy policy in July and made it less scary for users. Same goes for two of the children's toys, the Tamagotchi Uni virtual pet and Artie 3000 Coding Robot. Both of them keep the information about their users private and do not share them anywhere.

So, if you are going to give your loved one some tech device, make sure to read its privacy policy. Reading the privacy policy before accepting it is important to safeguard your private information.


Photo: DIW / AI-Gen

Read next: Google Maps Updated With Smart Routes, Easier Group Planning and Emoji
by Arooj Ahmed via Digital Information World

Microsoft Bing advances in search with AI-powered summaries

Bing's latest update is pretty cool. They're now using the fancy AI technology, specifically GPT4, to make searching online a bit easier for all of us. It's like having a smart assistant who quickly figures out what you're looking for and then points you in the right direction.

So, how does this work in real life? Let's say you're curious about finding alternatives to ChatGPT. You do a quick Bing search, and when you look at the search results, there's this little dropdown arrow next to the title links. Clicking on it reveals something new: a caption labeled "AI-Generated." This is Bing showing off its new trick.

Behind the scenes, Bing uses GPT-4 to sift through the maze of the internet based on your search. It picks out the important bits from web pages and turns them into neat and nifty summaries. These aren't just bland, copied-and-pasted lines from websites; they're more like smart, concise summaries tailored to what you're searching for.

The best part? These AI captions change depending on what you're asking. They're not just repeating webpage text verbatim. Bing's been pretty smart about making sure these snippets are both high-quality and accurate, even if they're not a direct copy from the sites.

If you're not into this AI caption thing, no worries. Bing's got you covered. You can turn them off with some simple tags like NOCACHE or NOARCHIVE. Bing's also cool with other tags like MAXSNIPPET and NOSNIPPET. They've got all the details in their guide on meta tags.

Now, why should you care about all this? Well, if you're managing a website or keeping an eye on how people find your content online, these AI captions could change the game. They might affect how often people click through to your site from Bing. It's a good idea to peek at your website's performance in Bing Webmaster Tools now and then. This way, you can see if these AI summaries are making a difference and tweak your strategy if needed.

It appears, Bing is rolling with the times, and it's worth keeping an eye on these new features. They could really change how we all use search engines.

Read next: AI's Data Dilemma Shows Less Can Be More
by Irfan Ahmad via Digital Information World

At Least 50% of Google Play Store Revenue Comes From Samsung Devices

If there is one company out there that has any chance whatsoever of competing with Apple as the foremost supplier of smartphones in the world, it’s Samsung. The only thing holding the Korean electronics juggernaut back is that it doesn’t have its own proprietary operating system, but in spite of the fact that this is the case, it is exerting enormous influence on the Android ecosystem with all things having been considered and taken into account.

The lawsuit that has been ongoing between Epic and Google has revealed some keen insights into the importance of Samsung at least in terms of Play Store revenue. With all of that having been said and now out of the way, it is important to note that the lawsuit revealed that Samsung accounted for at least half, if not more, of Play Store revenue back in 2019.

This comes from Google’s Vice President for Partnerships, James Kolotouros, and it’s also just the tip of the iceberg. Google agreed to pay $8 billion to Samsung over four years in order to make Play Store their default, along with Google Search and Google Assistant. This came after Google abandoned their intention to stop Samsung from placing Galaxy Store front and center. The carrot approach appears to have worked better than the stick, although it also meant that Google relied on Samsung more than might have been the case otherwise.

This also seems to suggest that Google is favoring Samsung over other OEMs. They received $2.9 billion in total back in 2020, whereas Samsung received $2 billion all on its own. Manufacturers also get a cut of Play Store revenue in order to prioritize the store, which might indicate an anti-competitive bias on Google’s part.

Manufacturers that play ball receive 20% of net basic ad revenue, 10% of net optimized ad revenue, 5% of net optimized Play transaction revenue, 15% of net premier ad revenue as well as 20% of net Play transaction revenue. The premium tier of this revenue split gives a full share of the revenue if OEMs pre-install apps offered by Google and avoid including those offered by its competitors.

It will be interesting to see where things go from here on out. Google’s relationship with OEMs is rapidly getting demystified, and many startling revelations are coming to the surface.

Photo: DIW / AI

Read next: Google News is Saturated by AI Plagiarism
by Zia Muhammad via Digital Information World

YouTube Looks To Get Ahead Of The AI Wave By Implementing Labels On AI-Content

Popular video-sharing app YouTube is trying its best to get ahead of others in the AI race. And if that means pulling out all the stops in terms of introducing features that inform viewers about AI content, then so be it.

The company published a new blog post that spoke about its plans to keep users updated about when they’re going to run into synthetic videos through the launch of descriptive tags. And all creators must abide by this requirement when producing content on the app.

The feature will come into effect in the next few months and that means disclosing AI-related material is now a major deal on YouTube. Anyone who has edited videos using AI software must also mention which AI tools were used throughout the process so others are aware.

While YouTube claims that any kind of animation generated through AI technology would be fine, a depiction featuring real individuals and events would now have to get tagged, sources added.

If in any case, a creator opts to bypass this rule, they would be penalized and the punishment could be as harsh as their removal from the platform complete suspension from all sorts of monetization activities, or simply having their account suspended if things are on the verge of being extreme.

From the looks of it, this means the severity would be linked to someone detecting the content featured AI but the creator failed to disclose it. Whatever the case may be, one thing is for sure. There would always be this consistent threat of removal arising and that’s a huge deal in terms of persuading many creators to do the right thing and avoid promoting any kind of AI-themed fakes.

While we agree it won’t end up deterring those channels that receive funding from foreign locations every year. But we do feel this is a step in the right direction in terms of giving users greater transparency when it comes to encountering AI-based content online.

The app also mentioned how it would take one extra step forward and add tags to content that’s been created using AI tools.

Additionally, we are also seeing the app include requests for the removal of any content that may have been used by AI software but failed to provide credit to the real owners of the material. And musicians are seemingly included in this ordeal.

We find this to be a unique aspect, especially for musicians as YouTube is reportedly aligning with plenty of record labels to produce more music using generative AI tools but with restrictions.

• Also read: Generative AI Receives Under 1% of Enterprise Investment, Traditional AI Hits 18%

YouTube has always played an integral role when it comes to discovering music through the app. And that has been a major concern for the platform. Therefore, we can see how this latest process will give rise to another huge step in terms of forcing rules for any kind of violation arising in the future.

But the app is not the first of its kind to take on this major step. We’re seeing tech giants Meta as well as TikTok also put forward tags for any kind of content produced using AI technology. And they’ve been mentioned for so long that the whole idea has to do with greater transparency and also an added dose of clarity for its viewers.

Many experts feel this will become even more integral in the next few years as more and more tools in the domain of AI continue to evolve.

Read next: Your Data at Risk — The Dark Side of Real-Time Bidding in Digital Advertising
by Dr. Hura Anwar via Digital Information World

X struggles with rampant hate speech in Gaza-Israel conflict; CCDH report exposes major moderation flaws

In recent weeks, X, the platform we all used to call Twitter, has become a hot topic, and not just for the latest viral trends. With tensions boiling over in Gaza-Israel, there's this growing buzz that X isn't exactly acing its own hate speech rules. The Center for Countering Digital Hate (CCDH) put out a report that's really turning heads.

Here's the scoop: the CCDH zeroes in on 200 tweets about the Israel-Gaza conflict. And we're not talking just any tweets; these are the kind that make you stop scrolling and think, "Wait, is this for real?" Packed with racist jabs and the sort of images that make your stomach turn. The CCDH flags these, expecting X to step up and take them down. But get this: a week rolls by, and 196 of these tweets are still out there in the wild, seen by who knows how many people.

And then there's Elon Musk's X. It's like he's trying to juggle free speech and moderation without dropping either. But with a whopping 43% of these controversial tweets from verified accounts, people are starting to ask if that blue checkmark on X might as well be a 'get out of jail free' card.

The CCDH’s deep dive reveals that just one account out of the 101 they flagged got suspended, and two others got a time-out (locked, in X lingo). And here's a number that'll make you blink – these posts from flagged accounts? They racked up a staggering 24 million views.

European Commissioner Thierry Breton's also chiming in, raising red flags about the uptick in iffy content on X. But Linda Yaccarino, who's now steering the X ship, is like, "We've got this" – talking up how they're cracking down on extremist accounts and slapping warning labels on content that's a bit suspect.

Musk’s entrance to X was like a scene from an action movie – full of drama. He chops the moderation team and rolls out the red carpet for accounts that had been previously shown the door. It’s sparked a ton of chatter. After all, it’s a fine line between letting folks speak their minds and not letting the platform turn into the Wild West.

So, yeah, X’s definitely got its hands full. Trying to find that sweet spot between upholding free speech and keeping a lid on hate speech and misinformation is no joke. It's like walking a tightrope, and everyone's got their eyes peeled, waiting to see X's next move.

Some FAQ about dealing with Hate Speech on Social Media:

Q: How can I recognize hate speech on social media?

A: Hate speech includes content such as racist slurs, dehumanization, and hateful imagery. Being able to identify these elements is essential in addressing hate speech.

Q: What should I do if I encounter hate speech on platforms like X?

A: Use the platform's reporting tools to report any hate speech. While not always immediately effective, reporting is crucial for content moderation.

Q: Is it advisable to engage with accounts posting hate speech?

A: If you choose to engage, do so cautiously and critically. Use reliable information to counter any misinformation in the hate speech.

Q: What role can I play in educating others about hate speech?

A: Inform and educate your online community about the impact of hate speech and the importance of respectful communication.

Q: How can I support better content moderation on social media?

A: Back organizations and initiatives that aim to hold social media platforms accountable for their content moderation practices.

Q: What steps should I take if I am a target of hate speech?

A: Protect yourself by using platform tools like blocking or muting accounts and adjusting privacy settings as needed.

Q: Why is it important to stay updated on social media policies?

A: Staying informed about platform policies helps you understand your rights and the platform’s responsibilities, especially regarding content moderation.

Q: Can user feedback influence social media platform policies?

A: Yes, advocating for stronger content moderation policies and providing feedback can influence how platforms evolve their guidelines.

Q: What is the importance of upholding personal moral standards on social media?

A: Maintaining ethical and moral standards in online interactions is vital. Being a role model for respectful communication helps influence the digital environment positively.

Read next: Sam Altman Predicts Relief From the High-end Chip Shortage, Impacting AI Development, by Next Year

by Irfan Ahmad via Digital Information World