AI Controversies and Innovations

Google’s Odd Pizza Glue Recommendation

In an amusing twist of AI-driven search results, Google recently suggested an unexpected pizza topping: glue. This bizarre recommendation stemmed from a viral moment when internet humorist Katie Notopoulos humorously experimented with adding glue to pizza, only to have Google’s AI mistakenly use this as a credible culinary suggestion. The search engine, pulling from current internet chatter, advised adding “an eighth of a cup” of glue to pizza—a clear mix-up in AI content generation.
This incident highlights the inherent challenges in AI-driven search technologies, which rely heavily on data from the web to generate answers. Such systems can inadvertently propagate errors or jokes from the internet into seemingly legitimate advice. It’s reminiscent of the early days of “Google bombing,” where users manipulated search rankings to associate phrases like “miserable failure” with President George W. Bush, showcasing how repetitive associations can skew search results.
The Verge confirmed the glue-pizza result through direct queries, adding a layer of authenticity—and concern—to the AI’s flawed learning process. This isn’t an isolated case. Even straightforward questions about Google’s own products, like taking screenshots in Chrome’s Incognito mode, yielded incorrect answers, underscoring the limitations of AI in parsing accurate information from complex data sets.
Comparatively, Perplexity.AI handled the same glue query with a sensible response, advising against putting any glue on pizza due to its toxicity. This disparity in AI responses underscores the variability in AI systems’ ability to discern and reject nonsensical data.
These humorous and sometimes frustrating misfires by AI search engines not only highlight the ongoing challenges in AI development but also emphasize the critical need for advanced algorithms capable of distinguishing between legitimate content and internet noise. As AI continues to evolve, the integration of more robust error-checking mechanisms and context-aware processing will be key to avoiding such entertaining yet misleading outcomes. Meanwhile, users remain part of an unintended global experiment in teaching AI about the world—with all its quirks and confusions. SOURCE 

Elon Musk Ends Lawsuit Against OpenAI, Altman

Elon Musk has retracted his lawsuit against OpenAI and its co-founders, Sam Altman and Greg Brockman, filed in California state court. The case was dismissed without prejudice, allowing for potential refiling, though no reasons for the withdrawal were disclosed in the court documents.
In February, Musk accused OpenAI, where Altman serves as CEO and Brockman as president, of breaching contract and fiduciary duties, stemming from a shift in the company’s mission. Originally established with the goal of developing artificial general intelligence (AGI) to benefit humanity, Musk contended that OpenAI had diverged from this path, becoming a profit-driven entity under significant influence from Microsoft, its principal shareholder.
This legal action highlighted Musk’s foundational role in OpenAI, which has gained prominence, partly due to the success of its AI product, ChatGPT. OpenAI was ranked first on CNBC’s Disruptor 50 list in 2023, illustrating its impact in the tech industry.
Legal experts had previously expressed scepticism about the solidity of Musk’s case, pointing out the absence of a formal, signed agreement at the core of his claims. The lawsuit was seen by some as more of a public relations move than a grounded legal battle, with experts questioning its substantial legal basis.
Besides his legal pursuits, Musk has been active in the AI sector with his start-up, xAI, which recently secured a $6 billion Series B funding round from notable investors like Andreessen Horowitz, Sequoia Capital, and Fidelity Management & Research Company. xAI aims to explore profound questions about the universe and released a chatbot named Grok, inspired by “The Hitchhiker’s Guide to the Galaxy,” showcasing real-time internet knowledge and two months of preparatory training.
This development underscores the intensifying competition in the AI field and highlights Musk’s ongoing interest in influencing the AI narrative, both legally and commercially. SOURCE 

Microsoft Removing Copilot GPT Builder from Copilot Pro Subscription

Microsoft is removing Copilot GPT Builder, a feature that allows users to create custom versions of Copilot GPT for specific tasks, from its Copilot Pro subscription service. This move comes into effect on July 10, 2024, and applies to both existing and newly created custom GPTs.

What is Copilot GPT Builder?

Copilot GPT Builder is a feature exclusive to Copilot Pro subscribers. It allows users to create custom AI models tailored to their specific needs. These custom models can be trained on documents, external or internal sources, and user-defined instructions. This personalization aspect makes Copilot GPT Builder more advanced than the default Copilot experience.

Why is Microsoft Removing Copilot GPT Builder?

Microsoft has not provided a clear reason for removing Copilot GPT Builder. However, there are a couple of possibilities:

• Cost-Cutting Measure: Copilot GPT Builder might be a computationally expensive feature to maintain for a consumer-focused product. By removing it, Microsoft could be aiming to reduce costs associated with Copilot Pro.
• Low Usage: It’s possible that user adoption of Copilot GPT Builder wasn’t as high as Microsoft anticipated. In this case, removing the feature could be a way to streamline the product and focus on more widely used functionalities.

What are the Implications for Copilot Pro Users?

The removal of Copilot GPT Builder reduces the value proposition of Copilot Pro for some users. Existing custom GPTs will be inaccessible after July 10th. Users who relied on these custom models will need to find alternative solutions.

What are Copilot Pro’s Remaining Benefits?

Despite the removal of Copilot GPT Builder, Copilot Pro still offers several advantages over the free Copilot experience:

• Access to advanced Copilot models during peak usage times.
• Ability to generate images with Copilot.

Alternatives to Copilot Pro

With the loss of Copilot GPT Builder, users who require custom AI models might consider alternative services like ChatGPT Plus, which offers similar functionality.

Microsoft’s Future Plans for Copilot

While Copilot GPT Builder is being removed from the consumer version, Microsoft remains committed to Copilot technology. The company is exploring new ways to extend Copilot’s capabilities, but details are not yet available. Additionally, Microsoft will continue to offer GPTs and GPT Builder functionalities to commercial and enterprise customers. SOURCE 

Apple Unveils Approach to Generative AI at WWDC 2024

Apple shed light on its plans for generative AI at the Worldwide Developers Conference (WWDC) 2024. This marks a significant development, especially considering the contrasting approaches taken by tech giants like Samsung. Here’s a breakdown of the key takeaways:

• Apple’s Focus: Apple is prioritising user privacy with its generative AI implementation. This is a potential distinction from competitors who might prioritise different aspects.
• Integration Strategy: Apple plans to deeply integrate generative AI features throughout its ecosystem, encompassing iPhones, iPads, and Macs. This suggests a cohesive user experience across devices.
• Samsung’s Approach (for Context): It’s unclear how Samsung has approached generative AI so far, but mentioning the contrast implies Apple might be taking a different direction. This could involve a more cautious or privacy-centric approach compared to Samsung.

Overall, Apple’s announcement highlights its commitment to generative AI technology, while prioritising user privacy and offering a well-integrated experience across its product range. This stands in contrast to some competitors, whose strategies might differ.
While both Apple and Samsung acknowledge the transformative potential of generative AI for smartphones, their approaches unveiled at WWDC 2024 showcase distinct philosophies.

Apple Intelligence:

• Focuses on unifying the user experience across Apple devices (iPhone, iPad, Mac) through AI features.
• Offers functionalities like message rewriting, image generation, and enhanced Siri capabilities for complex questions.
• Aims to leverage AI for creating connections between different apps used by the user.

Samsung Galaxy AI:

• Primarily targets specific tasks within individual apps, such as photo editing or language translation.
• Overlaps with Apple Intelligence in areas like text summarization, proofreading, and note rewriting.

Key Differences:

• Integration: Apple prioritises a holistic AI experience across devices, while Samsung focuses on AI within individual apps.
• Functionality: Apple emphasises user interaction and task completion across platforms, whereas Samsung leans towards specific tools within apps.

The Future of AI in Smartphones:

This is just the beginning. Both Apple and Samsung hint at further advancements in their AI offerings. Evaluating the user experience and effectiveness of these features will require real-world testing.

Overall Impact:

Apple and Samsung’s contrasting approaches set the stage for how generative AI might evolve in smartphones. Their strategies will likely influence the direction of AI integration in mobile devices across the industry.

Apple Unveils Personalized Generative AI Strategy with Apple Intelligence

At WWDC 2024, Apple entered the generative AI space with a distinct approach called “Apple Intelligence.” Here’s a breakdown of the key takeaways:

Focus on Personalization:

• Apple prioritises user context and personal data. AI features aim to understand and leverage your messages, files, photos, and habits for a more intuitive experience.
• Craig Federighi, Apple’s senior vice president of software engineering, emphasises AI integration within existing user experiences.

Enhanced Siri Capabilities:

• The new Siri boasts improved device search functionalities. Siri can now locate information across messages, emails, notes, photos, and calendar events.
• Imagine forgetting where a recipe from a friend is stored; Siri can search through various apps to find it.
• Siri can leverage personal data for actions beyond search. For instance, Siri can potentially fill out forms using information from your driver’s licence photo.

Going Beyond Search:

• Apple envisions Siri as a proactive assistant, not just a search tool. Imagine asking Siri to “edit a photo of Stacy in New York with her pink coat and add it to a note.” Siri, understanding context from your Photos app, can complete the task through voice commands.

AI in Photos App:

• Photos gets a personalised AI boost. Users can create custom photo memory movies with prompts like “Leo trick-or-treating with a spooky vibe” or “Everything we ate in Japan.”

Apple’s Approach in Context:

• While Apple’s focus on personal data is unique, the core concept of referencing information across apps isn’t entirely new.
• Microsoft’s Copilot assistant and Google’s Gemini virtual helper exhibit similar functionalities. It’s likely Samsung will integrate similar features soon, given their partnership with Google.

The Future of AI Assistants:

• The focus on personalised AI and task completion suggests a future where virtual assistants like Siri can handle more complex errands, potentially inspired by Google’s vision of AI agents returning items for you.

Apple’s “Apple Intelligence” signifies a distinct approach to generative AI, prioritising user context and leveraging personal data to create a seamless, personalised user experience. This approach sets the stage for an exciting future of AI integration within smartphones, with competitors likely offering their own variations. SOURCE 

Samsung vs. Apple: Contrasting AI Strategies on Smartphones

Samsung prioritises communication and productivity in its smartphone AI strategy, exemplified by:

• Real-time call translation within the native phone app.
• Text rewriting, proofreading, and full conversation translation across languages.
• Generative editing for photos, including object removal, resizing, and movement.
• Video slow-motion previews, note summarization and translation, and “Circle to Search” for on-screen information.

This focus on communication and content creation differs from Apple’s WWDC presentation, which lacked emphasis on translation features.

Convergence Over Time?

Interestingly, Apple’s Siri advancements resemble Samsung’s original vision for Bixby (phone navigation and app interaction) rather than typical virtual assistant functionalities. This suggests potential overlap in future iterations of both companies’ AI offerings.

The Future of Smartphone AI

Both Samsung (with a generative AI upgrade for Bixby) and Apple (with likely expansion of Apple Intelligence) are committed to improving smartphone AI for user productivity. The coming months will be interesting as Samsung unveils new foldable phones and Apple potentially releases new iPhones, showcasing their latest advancements in this domain. SOURCE 

Conclusion

Apple and Samsung are taking different approaches to integrating generative AI into smartphones. Apple prioritises user privacy and a holistic user experience across devices, while Samsung focuses on specific tasks within individual apps. Both companies are committed to improving smartphone AI for user productivity, and their advancements will shape the future of this technology. Join us at Arcot Group, where innovation meets excellence. Discover our latest projects and see how you can be part of a future driven by technology and leadership. Click here to learn more and get involved.