Saturday, June 22, 2024

Is AI Killing Creativity? OpenAI's Shocking Revelation Sparks Outrage!

OpenAI often professes its mission to develop a superintelligent computer system that benefits humanity. However, it frequently stumbles in its communication, portraying AI in a manner that seems more likely to alienate rather than unite people.

During a discussion about AI at Dartmouth University, hosted by the engineering department and featuring Dartmouth Trustee Jeffrey Blackburn, OpenAI's CTO Mira Murati made a remark about the impact of AI on creative professions. She suggested that some creative jobs might disappear because the content produced was not of high quality, implying that such roles perhaps should not have existed in the first place. Murati added that AI could enhance intelligence when used as a tool for education and creativity.

For those skeptical of generative AI who view it as largely built on the illusion of originality, Murati's statements are especially aggravating. Her comment comes across as dismissive, insinuating that certain professionals do not deserve their jobs in the AI era. This sentiment is infuriating not only to creative professionals but to anyone who values their work and effort.

Image: DIW-Aigen

Previously, Murati seemed uncomfortable when questioned about whether OpenAI's Sora video tool was trained using YouTube videos. This incident contributes to the growing perception that OpenAI is antagonistic toward creative industries. There is a notable lack of trust in the authority of OpenAI CEO Sam Altman and his team to decide whose job is valuable, especially after Altman’s controversial comments about AI potentially replacing "median humans."

Many geeks who are outside the AI development sphere, hold the belief that human potential is unique and invaluable. Every individual deserves an opportunity to realize their capabilities fully. Contrary to this belief, OpenAI appears to undermine these values. This is the same company that once appropriated an actress's voice without consent. Ed Zitron, in one of his recent newsletters, highlighted the dangers of not holding such companies accountable. He warned that failing to scrutinize their practices would harm creatives whose work is being appropriated. Zitron also criticized the corporate tendency to integrate AI into products to satisfy investors, which, he argued, detracts from the tech ecosystem’s value and enriches a few at the expense of many.



This post was published using the help of AI.

Read next: Apple Delays EU Launch of Apple Intelligence Due to Antitrust Rules
by Web Desk via Digital Information World

Apple Delays EU Launch of Apple Intelligence Due to Antitrust Rules

Apple won’t be releasing Apple Intelligence and three other features in the EU in 2024. This is because of privacy concerns around the Digital Markets Act (DMA). The regulation requires services to be interoperable across platforms to promote competition and reduce the power of big companies like Apple.

The company says Apple Intelligence, iPhone Mirroring and SharePlay screen sharing will not be available to EU users. The interoperability requirements of the DMA pose a risk to user privacy and data security and Apple won’t compromise on that.

The DMA was introduced in 2023 to address the power of big tech companies like Amazon, Alphabet, Meta, Microsoft and ByteDance (TikTok’s parent company). It requires fundamental features to work seamlessly across different devices and ecosystems. Although it mainly affects iPhones and iPads, Macs are also impacted because of iPhone Mirroring which allows iPhone screens to be mirrored on Mac screens.

Users may be disappointed Apple Intelligence won’t be available. This AI product had features like proofreading, custom emojis called Genmoji, searching specific messages, summarizing and transcribing calls, and highlighting priority notifications. Apple also announced a partnership with OpenAI and plans to add more models to the platform.

Despite the news Apple’s stock didn’t move much. In 2023 Apple made $94.3 billion in Europe which is almost a quarter of its global net sales. Apple also said Apple Intelligence won’t be available in Greater China which made $72.6 billion of its 2023 sales.

Apple will continue to work with the EU to find a way to deliver these features to EU users.

Image: DIW-Aigen

Read next: New Benchmark Shows AI Agents Struggling with Real-World Tasks
by Asim BN via Digital Information World

Friday, June 21, 2024

New Benchmark Shows AI Agents Struggling with Real-World Tasks

A customer experience AI startup, Sierra, has developed a new benchmark that helps in evaluating the performance of AI chatbot agents. The benchmark is named TAU-bench and is evaluated by having conversations with LLM-stimulated users while doing complex tasks. The results show that AI agents which are made with simple LLMs are not able to perform simple tasks. This means that companies need more advanced AI agents for work.

Sierra’s head of research, Karthik Narasimhan, says that Sierra’s benchmark is helping real-world users evaluate the performance and reliability of the AI agents which is important if you want the AI agents to work in a real world setting. He also added that many other benchmarks like SWE-bench, Agentbench and WebArena have also been created for the same purpose but they are not successful in working to their full extent. They are only able to evaluate a single round of agent-human interaction, without answering about more dynamics. This makes them less reliable and adaptable.

As other benchmarks were experiencing some issues, Sierra addressed those issues with TAU-bench. It represented three requirements for the benchmark– the agents should interact smoothly in the real world settings, agents should follow rules and policies given for the task and agents should be reliable so companies can work without having to worry about their results. Many tasks were given to TAU-bench like working on real databases to different APIs and other complex tasks that required it to have conversations with the agents. Every assignment given the agent was about retaining information, performing complex tasks and communicating through real conversation.

The four main features of Sierra’s benchmark after experimenting with it were TAU-benchmark can do realistic dialog, can perform open ended and diverse tasks, do faithful objective evaluation and provide a modular framework. TAU was tested using 12 popular LLMs including GPT-4, Claude-3, Gemini and Llama. All the agents performed poorly, including ChatGPT-4 which got less than 50% average success rate in all domains.



Read next: The 5 Eyes Alliance Can Spy On You Wherever You Are, Here's What You Can Do to Protect Yourself
by Arooj Ahmed via Digital Information World

The 5 Eyes Alliance Can Spy On You Wherever You Are, Here's What You Can Do to Protect Yourself

When Edward Snowden revealed that the NSA was secretly spying on American citizens, people began to realize the inherent fragility of the world we are all living in. Snowden’s expose shed some light on the murky, interconnected intelligence networks around the world. Most countries have comprehensive intelligence networks, and while they are legally prohibited from spying on their own citizens, collaborations between these agencies provide a bit of a loophole.

Five major nations have banded together to form the Five Eyes alliance. These countries are the US, Canda, the UK, Australia and New Zealand. They share several similarities such as being former British colonies, Anglophone nations as well as having overlapping geopolitical interests with all things having been considered and taken into account.

Another common thread between these nations is that they were all on the same side during World War 2. hence, it is unsurprising that they decided to join forces in 1946. With the USSR emerging as another superpower to the East, challenging the neoliberal world order that the US and its allies were trying to cement, the Five Eyes alliance attempted to keep the rising red tide at bay.

With all of that having been said and now out of the way, it is important to note that the NSA is the unofficial leader of the Five Eyes alliance. Signatories to the agreement are able to access vast quantities of NSA data, and they return the favor by sharing secret information that they may be in possession of.

The other four countries, namely Australia, New Zealand, Canada and the UK are referred to as second parties. They have the most unrestricted access to NSA data, but it bears mentioning that several other nations have signed on as third parties.

NATO members are all automatically included in the third party list, as are other close US allies such as South Korea. They have more restricted access to the NSA’s database, and they also lack other privileges.

Five Eye members are not allowed to spy on each others citizens. This privilege does not extend to third party signatories, thereby running the risk of their confidential secrets falling into the wrong hands. While the general assumption is that Five Eye nations don’t spy on each other, there is no way to prove this. After all, the NSA shouldn’t have been spying on American citizens either, but this did not stop them from trying to keep tabs on people that had not even committed a crime to begin with.

Apart from the Five Eyes alliance, there are a couple other intelligence networks that include the same five nations along with a few others. First and foremost, the Nine Eyes alliance widens the network to include non-English speaking nations whose goals and interests align with those of the US and its closest allies.

Denmark and Norway joined the Nine Eyes network from Scandinavia, with the Netherlands and France also joining from Central and Western Europe. One thing that bears mentioning is that the Nine Eyes alliance is not as official as the Five Eyes alliance. It comprises third party signatories that have certain exclusive privileges, so it is more of an unofficial agreement that does not possess any type of legal backing with all things having been considered and taken into account.

Now, there is another labyrinthine intelligence network that hardly anyone knows about, namely the 14 Eyes Alliance. One again, the Five Eyes and their formidable intelligence apparatus as at the heart of this alliance, as are the four nations that are included in the nine eyes, but five other major players are thrown into the mix as well.

Each and every one of the members of the additional five are from Europe, including Germany, Sweden, Belgium, Spain and Italy. This clearly shows that the Western powers of the world have extremely close ties, since only North American and European nations are official signatories in any capacity whatsoever.

In spite of the fact that this is the case, a few other nations that go beyond European borders are also seeing an increased level of prominence in global intelligence affairs. We have already mentioned South Korea’s inclusion as a third party, with other powerful East Asian nations such as Japan and Singapore also joining it in that regard. Israel is another unofficial member of this exclusive club, with the US frequently propping it up through the provision of military aid.

India and Thailand are also seeing quite a bit of prominence on this front. India is rapidly growing, and it may soon become the largest economy in Asia and perhaps even the entire world. It may be the sole nation that can compete with China for Asian dominance, so Western powers are attempting to bring it into the fold.

China has its own rival intelligence network that it has been building for quite some time now. The rising East Asian great power could potentially end up becoming a superpower on par with the US itself, and it has a range of regional powers that is relying on.

The name of China’s alliance is the Shanghai Cooperation Organization. It includes countries in the Russosphere such as Kazakhstan, Kyrgyzstan, Tajikistan, Uzbekistan as well as Russia itself, along with Pakistan which is notable considering it is a nuclear power. Interestingly, even India is a part of this alliance despite the inclusion of its greatest political rival. This makes India the only nation that is able to toe the line and not pick sides in the global war for domination.

While it might seem like these nations are collaborating in an entirely above board manner, this couldn’t be further from the truth. We only know about the informal 14 Eyes Coalition due to a leak, which just goes to show that governments would much prefer to keep these types of things under wraps.

So, what does this mean for the average consumer? Well, for starters, it means that we all need to start being a lot more careful about the manner in which our data is transferred and stored. Given how frequently these nations share data with each other, consumers must be informed about how they can protect themselves.

According to the Electronic Frontier Foundation, these countries can easily bypass what they are referring to as the lowest common privacy denominator. Mass surveillance is at an all time high, and it is up to us to keep ourselves safe from prying eyes.

Many of these countries are trying to codify their attempts at mass surveillance. If any of this legislation ends up getting passed through the requisite legislative bodies, we may end up living in a society where no one would be able to keep a secret from the government.

There are several ways in which we can all keep intrusions at bay. First and foremost, it is imperative that we start encrypting all of our data. The single best way to go about doing that is by using some kind of a VPN.

A Virtual Private Network sends your data through various digital servers, thereby obscuring your footprint from anyone that may be trying to spy on you. Your internet service provider can monitor your online activities and share it with intelligence agencies if they present a warrant, but with a VPN, they would not have this data to begin with unless they procure it from the VPN provider.

Unfortunately, many VPNs are located in countries that are members of the Five Eyes network. While they might not necessarily give your data up, the countries that participate in this agreement are known for stopping at nothing to get what they need.

A far superior solution would be to use a VPN that is based in a nation that is not a signatory of any intelligence sharing agreements whatsoever. Now, with so many countries freely sharing intelligence, is there really a country out there that can stand out in any capacity?

ProtonVPN might be the answer, since its servers are housed in Switzerland which is famous for being the most privacy friendly country in the world. Switzerland has a long history of being neutral in all conflicts. That can be useful because of the fact that this is the sort of thing that could potentially end up keeping your data completely and utterly safe, no matter how powerful the nation demanding it currently happens to be.

The main benefit of Proton VPN is that it uses something called Perfect Forward Secrecy. This generates an entirely new key for your various browsing sessions, thereby making it so that your other sessions will be perfectly safe even if one of them end up being compromised to any extent whatsoever.

Apart from encrypting your data with a VPN, you should also consider using something similar for your emails. The Five Eyes alliance uses something called SIGINT which refers to the interception of transmissions between two points. It was used almost exclusively in warfare, but following the aftermath of the Cold War, it was more and more frequently used to spy on regular everyday citizens.

It stands to reason that your own emails are passing through their filters, but with Proton VPN, these emails will be encrypted. Even if they end up getting intercepted, the information contained within them would not be readable. Only you as well as the intended recipient would be able to read whatever it contains.

Some might also recommend that you use privacy conscious browsers. They don’t collect your data similar to Google Chrome and others, thereby giving you an added level of security that you would certainly be grateful for as the global surveillance network becomes ever more powerful.

ProtonVPN yet again offers something quite useful in this regard. Their email service uses a zero trust protocol, which is a level of encryption so advanced that even the company itself would not be able to crack it. On the off chance that an intelligence agency is successfully able to convince Proton Mail to hand over your data, they can simply state that they are unable to do that even if they wanted to.

This allows you to depend on technology rather than an institution and its ethics and morality. No matter how trustworthy a company seems, things like zero trust encryptions can ensure that your privacy will remain unassailable for the foreseeable future.

It would also be extremely useful for you to take a look at messaging apps that offer encryption. Signal and Telegram both stand out, although even something like WhatsApp might get the job done thanks to its end to end encryption.

The global agenda is turning mass surveillance into a foregone conclusion. Even if you don’t live in any of the countries that were mentioned in the Fourteen Eyes Alliance or the Shanghai Cooperation Organization, you might still find yourself under surveillance without any knowledge of what is going on.

The use of ProtonVPN, Proton Mail and other types of services in a similar vein could prove to be the last line of defense. The governments of the world clearly don’t care about their citizens, rather they only seem to care about furthering their own ambitions. A decentralized, encrypted future is the only way forward, at least for those that don’t want to be under the microscope at all times.


Country Five Eyes Nine Eyes Fourteen Eyes Other
United Kingdom
United States
Australia
Canada
New Zealand
Denmark
Netherlands
France
Norway
Germany
Belgium
Spain
Sweden
Italy
Israel
Japan
Singapore
South Korea

Read next: Which Global Airports Have The Fastest Wi-Fi Speeds? The Answer Might Surprise You
by Zia Muhammad via Digital Information World

Anthropic’s New Powerful Claude 3.5 Sonnet AI Model Set To Provide Serious Competition To Other AI Archrivals

The AI competition is heating up as more generative AI models pop up at the speed of light.

The recent entry seems to be from OpenAI rival Anthropic who just rolled out a new and powerful offering called Claude 3.5 Sonnet. However, experts claim the upcoming release might be more of an incremental step than a major leap in the forward direction.

The model can carry out analysis of text and images alongside text generation. And to date, it’s believed to be the company’s best offering so far.

When we look at the statistics, the Claude 3.5 Sonnet does a better job in terms of performance for this model than it would replace which is the Claude 3 Sonnet. It ends up beating the company’s former flagship model which is Claude 3 Opus.

While benchmarks are not usually dubbed to be the most useful means for determining progress and performance, it’s being used here and experts feel it might be slightly better than its older variant and greater in performance to OpenAI’s famous GPT 4o which is major news as it was tested against both of them.

Other than this new model, the company is rolling out another exciting workspace endeavor called Artifacts where users could edit and add to written content such as codes and documents that are produced by models from Anthropic. While it’s in the stage of preview for now, this will soon gain more features including ways for collaboration with bigger groups and storing knowledge bases for the future.

The company’s main goal right now seems to be efficiency. The latest Generative AI model is doing great in terms of performance, the company adds, and is understanding better when complex instructions are rolled out in its direction like humor.

The speed is another great factor worth mentioning. The model’s speed of rolling out prompt responses is touching the sky as it’s nearly twice as fast as the model from the past, the company adds.



Other points worth a mention include spectacular vision where photo analysis is done so well in one place while chart and graph interpretation is more accurate text transcription is also carried out from images that might not be so great like distorted graphics or those featuring artifacts.

So how come so much difference between it’s older model? The product head at Anthropic says major enhancements are arising due to changes in architecture and making use of new training data like AI-generated material which the company is yet to disclose what the real source is.

The secrets linked to training data here have to do with competition and also to prevent the company from getting targeted legally by others. Remember, there are plenty of competitors in the industry such as Amazon, Google, and even OpenAI. So getting the right source of training from copyrighted content belonging to others without giving compensation is really turning out to be a problem right now.



Read next: YouTube’s New App Tracking Transparency Prompt To Soon Ask iOS Users If They Want Personalized Ads
by Dr. Hura Anwar via Digital Information World

YouTube’s New App Tracking Transparency Prompt To Soon Ask iOS Users If They Want Personalized Ads

With the launch of iOS 14.5, we saw tech giant Apple force developers to inform users about an opt-in if they wished to track them down. But then in 2021, all apps from Google stopped indulging in certain kinds of tracking for compliance and also rid of that prompt altogether.

Now, the latest on this front is that popular video-sharing app YouTube is once again rolling out a new ATT prompt where users would first be asked to opt in to enable personalized tracking for ad purposes.

The app explained how it hopes to include the one-time permission feature where YouTube is giving users the chance to ‘allow’ the platform to track their activity via two options.

The first one says Allow and that means when selected, all activity from the app would be linked to that from non-Google platforms and pages to showcase personal ads and enable ad measurements. The user’s selection would apply to all on the app when turned on.

The second option says ‘Ask App Not to Track’ where if the setting isn’t turned on, no activity would be linked through non-Google platforms and pages for personalized ad sharing and measurements. This would give rise to a lower frequency of relevant ads and those that are repeated more often than others.

Such prompts aren’t going to be a part of the Premium subscribers and those having child accounts on the app would also not see this.

When you click on ‘Allow’, the platform claims users would expect to see better or more useful ads which were of greater quality. For instance, the ads popping up on the app could better display what they believe in or what their interests are aligned with.

On the other hand, the app says the feature will soon be up for grabs to all iPhone and iPad users who will now have the chance to be tracked in a more sophisticated manner which not only benefits the company but also is more aligned with the users’ interests.

Furthermore, the user can opt to switch that feature off whenever they want by going to their settings and then clicking on the privacy and security tracking tabs.

We don’t see how this setting could impact the linking of user activity via other apps from Google and its respective pages so it’s going to be welcomed in our opinion. What do you think?

Image: DIW-Aigen

Read next: Researchers Raise Alarm Against Instagram’s Algorithm For Serving Sexual Content To Young Teens
by Dr. Hura Anwar via Digital Information World

Researchers Raise Alarm Against Instagram’s Algorithm For Serving Sexual Content To Young Teens

A new study is unraveling the dark side of Instagram and its algorithm where alarming statistics proved how young teens are being recommended sexually charged content.

The recommendations feature creators in ‘barely there’ attire where underage viewers as young as 13 were a part of the audience.

The news arose in the latest study from WSJ and another Northeastern University which says it’s about time Meta took notice and serious action against the behavior while calls for banning such accounts were also made by activists noting down the issue.

On the first visit, researchers saw how young viewers’ recommendations were mostly from females dancing around in revealing clothing where their cleavage was completely visible.

The fact that viewers were seeing these posts more often meant that graphic content kept on becoming a part of the algorithm’s recommendations which many feel is just not okay. Most of the videos were a part of the app’s Reels feed and suggestions were mindblogging for some.

Other alarming content featured sex workers promising viewers nude pictures in their DMs. And that again was happening so frequently, the published study explained.

In another set of fixed tests carried out in June, the Journal began explaining how content after content became more graphic in nature including anal sex being openly promoted to those as young as 13, provided the viewer saw Reels related to women on their feed.

Seeing the algorithm go to this extent to serve content of this kind where females caressed their bodies and did everything from flashing private parts to more was just a lot for the authors to handle, adding how Instagram’s algorithm needed a big-time clean-up before matters became worse.
Meta has pushed back on those findings of the report, where the spokesperson brushed the matter off as just an experiment conducted that didn’t match up in real-time with what youngsters were being exposed to.

Meta says it has time and time again managed to reduce sensitive themed content linked to teens to such an extent that it’s at a bare minimum, completely eliminating what the research had to say.

The fact that this study was carried out for seven months is again alarming as Meta keeps denying how the results were just part of an artificial experiment and nothing more. While the test was certainly carried out to see how the app responded to illicit recommendations, it’s mind-boggling how so much is being ignored and for so long in that experiment.

Over the years, the app’s algorithm has been questioned in the past too but the tech giant keeps claiming it’s rolling out stricter measures to ensure underage users remain safe at all times on the app.

One example of that is those below the age of 16 are automatically blocked from viewing explicit content and recommendations.

Image: DIW-Aigen

Read next: 

Reddit Witnesses Massive Growth In Visibility On Google Search But It Comes With Implications

59% of Marketers Report New Business Opportunities from Content Marketing

• Which Global Airports Have The Fastest Wi-Fi Speeds? The Answer Might Surprise You
by Dr. Hura Anwar via Digital Information World