Ken Flott’s Post

Something to be aware of in our new age of technology.

LinkedIn has confirmed the use of personal user data to train its AI models, sparking concerns about user consent. The platform collects data on user/our activity, posts, language preferences, login frequency, and feedback provided. But remember, you can opt out of the AI tool on mobile and desktop by going to Settings, clicking “Data Privacy,” and selecting “Data for Generative AI Improvement.” Toggle the feature off to prevent further data usage, although data already captured will remain in use for AI training. #LinkedIn #AI #DataPrivacy

LinkedIn has a new practice of sharing your personal data to train AI - unless you specifically opt out. That includes your profile, your posts, and your videos. Without announcing it, LinkedIn apparently added a new data privacy setting last week that covers this, and they turned it on for everyone. If you want to opt out, in your LinkedIn account, open “Settings & Privacy.” Select “Data privacy” and turn off the option under “Data for generative AI improvement.” Be aware that turning this off will be not retroactive. LinkedIn has already begun training its AI with your content, and there’s no way to undo whatever they’ve already used.

The future of AI search is at stake. Multiple lawsuits are aiming to take down AI search engines because of their ability to retrieve and repackage information from third-party websites. While AI search offers a new, concise, way for people to understand the world around them, some are worried that, if this method becomes our new standard, it will disrupt our digital economy. Your thoughts? #AI #SearchEngines

You never search beyond Google. But Perplexity AI makes it old-fashioned: Because everyone's tired of Googling. ↳ You spend 15 minutes going from website to website. ↳ You try summarizing it manually & get rid of ads. So Why Perplexity AI is (Way) Better than Google (so many of you have asked for this post) The short version is: "It’s all about speed & precision." The long version? 1. Specialized Search Experience: ↳ Delivers precise and relevant outcomes. ↳ Understands natural language queries. 2. Tailored Results: ↳ Intuitive for specialized topics. ↳ Handles complex questions for focused answers. 3. Deep Research Capabilities: ↳ Provides context-aware insights. ↳ Ideal for research-intensive tasks. 4. Advanced AI Technologies: ↳ Utilizes ChatGPT-4o and Claude 3.5. ↳ Offers comprehensive summaries & direct footnotes. 5. User Interaction and Feedback Loop: ↳ Continuously fine-tunes results. ↳ Learns from user preferences and behaviors. 6. Innovations in Data Privacy: ↳ Uses differential privacy and federated learning. ↳ Ensures minimal collection of sensitive data. 7. Ad-Free Experience: ↳ Cleaner and more focused search results. ↳ No clutter from ads. 8. Standardized Interface: ↳ Consistent and user-friendly. ↳ No annoying pop-ups or hard-to-read layouts. 9. High Accuracy: ↳ Answers complex questions accurately. ↳ Proven performance in extended trials. 10. Potential to Revolutionize Search: ↳ Unique approach powered by advanced AI. ↳ Could challenge Google's dominance. Perplexity AI is the future of search. Do you agree? ♻️ Repost this if you do. PS: I wrote this post entirely with my AI, EasyGen. If you want a quick video demo, send me a message. The proof is on my top comment.

Perplexity makes search a lot more comfortable. But it also makes us lazy. It is good for a quick research, but it greatest asset is also its greatest disadvantage: you only get and see one answer, whereas the reality quite certainly is much more complex. Plus, regarding political topics, the overton window gets defined by Perplexity and its programmers. It conveys by far not t h e truth, but rather biased information, possibly with an agenda.

Yet another development from the March 2024 update. Google now mentions AI-content directly in the Quality Rater Guidelines: In section 4.7, the "Examples of Lowest Quality Pages" they added an example site that they flag for having low quality content. It cites how the publication utilized AI as a reason for having a low quality assessment: "The website terms of use states that 'some articles' are generated by artificial intelligence and may have errors or be out of date; there is no indication to which pages this statement applies. The information in this article is not trustworthy and is Lowest E-E-A-T" They also go onto to state that given the statements in the website terms the site's articles"should be considered untrustworthy and Lowest E-E-A-T" So in this example, it seems Google is taking into consideration how websites utilize and self-identify when generative AI is used to create content. Credit to Barry Schwartz & Natzir T. for finding this.

You can now ask a follow-up question using Brave Search's "Answer with AI." Introduced a week ago. I use this service regularly and will definitely be using the new follow-up feature. What I like about Brave Search? No login required, and they are committed to user privacy. Other than Brave Search, you may also want to try Brave browser, which I also use regularly. Links: Brave's announcement on LinkedIn about AI chat for follow-up questions: 🔗 https://lnkd.in/g9cYh9YF TechCrunch's article on the Brave Search new feature: 🔗 https://lnkd.in/gzxBzAsd Brave's blog post introducing their Privacy-focused AI Answer Engine: 🔗 https://lnkd.in/g455Xzva Try Brave Search: 🔗 https://search.brave.com/

Google’s AI Overview – a New Feature Facing Criticism 👇 Google's AI Overview, a new feature providing quick answer summaries in search results. It’s like Google doing the googling for you! It has been live for two weeks and is already facing criticism. Users are reporting that AI Overview has instances of providing inaccurate and misleading information – like adding 1/8 cup of nontoxic glue to pizza sauce to let the cheese stick! 🤭 Now this raises concerns about the reliability of its data sources and filtering processes. This isn’t actually Google’s first issue with its AI tools. Gemini, their generative AI tool, also faced issues in generating images. Anyway, the key takeaway here is that fact-checking remains crucial with any AI tool as they are still developing. ✅ We can always try AI overview, but use it with a critical eye, and see how it will evolve in the coming months. What are your thoughts? Does Google also need to rethink its AI approach to avoid falling behind? https://lnkd.in/g5X2FJng #google #aioverview #artificalintelligence

What do you need to know about #GenAI and retrieval-augmented generation? This blog breaks down: 🔵 Why #AI applications use RAG with their #LLMs. 🔵 What the potential impact for businesses is if LLM RAG targets their website. 🔵 How to block LLM RAG requests on websites and APIs. Learn more: https://lnkd.in/g_WEv2XJ DataDome