Everyday AI

Deepfakes are synthetic media in which a person in an existing image or video is replaced with someone else's likeness using AI algorithms, particularly deep learning techniques such as generative adversarial networks (GANs). Generative AI has ushered in a new era of media manipulation, with deepfake technology at the forefront.

In this post, I'm sharing a worry of mine about a potential threat to our constitutional “rule of law”. The rapid advancement of AI and deep learning technologies has ushered in an era where digital forgeries, referred to as "deepfakes", pose a major threat to the integrity of courtroom evidence.

By enabling robots to utilize cameras for visual training, MIT researchers are effectively granting AI systems a form of visual self-awareness. This development raises profound questions about the trajectory of AI and robotics, particularly regarding the concept of self-awareness, a trait long considered exclusive to biological organisms.

The AI Browser Wars are not a fad; instead, they represent a basic reordering of the internet’s value chain. The "answer first" paradigm is already eroding the traditional web “link first” traffic model, undermining the economics of content creation and forcing a radical rethink of SEO and digital advertising.

AI Sycophancy refers to the tendency of generative AI models to adapt their responses to align with the user’s views, even when those views are factually incorrect or ethically problematic. The term, borrowed from human behavior, describes a pattern where AI chatbots prioritize user approval over truthfulness or objectivity.

A key element of learning is “learning” how to study; learners of all ages struggle with developing good study habits. Knowt, founded in 2018, is a U.S.-based EdTech startup that leverages AI to revolutionize note-taking, study, and assessment. The platform’s core offering is its ability to automatically generate quizzes and study materials from user-uploaded notes, PDFs, and textbooks, using advanced natural language processing (NLP) algorithms.

The AI Assistant Economy refers to the ecosystem of technologies, businesses, and services centered around AI-powered virtual assistants. These assistants are software programs designed to perform tasks, provide information, and interact with users through natural language processing (NLP) and machine learning. AI assistants are no longer limited to personal use. They are now integrated into industries such as healthcare, finance, retail, and logistics. 

We are once again at the crossroads of another significant opportunity. The integration of AI in education is revolutionizing how students learn, teachers instruct, and institutions operate. Khanmigo for Education (KfE) is the latest brainchild of Salman Khan (Khan Academy). This post introduces Khanmigo.

At the Worldwide Developers Conference (WWDC) 2025, Apple made a groundbreaking announcement that could redefine the landscape of artificial intelligence (AI) in mobile and desktop applications. The company revealed that it is opening its Apple Intelligence Foundation Models Framework (FMF) to third-party developers, enabling them to integrate on-device AI capabilities into their apps.

The integration of BCIs (Brain-Computer Interfaces) with AI is reshaping the boundaries of human-machine interaction. AI enhances BCIs by processing and interpreting complex neural data, enabling more precise, adaptive, and efficient control of external devices. This convergence has the potential to revolutionize healthcare, augment human cognition, and redefine the interface between humans and technology.

This week, I thought we should discuss one of AI's darkest realities, which we all need to accept and do all we can to manage the consequences. Axios published remarks made during an interview with Dario Amodei, CEO of Anthropic, who shared a blunt, frightening warning for the US government. and all of us. Amodei shared that AI could wipe out half of all entry-level white-collar jobs, and spike unemployment to 10-20% in the next one to five years.

The Open Agentic Web refers to an evolution of the Internet in which autonomous, AI-powered agents act as intermediaries between users and digital services, operating across platforms, applications, and devices. These agents are not siloed within proprietary ecosystems but are interoperable, open-source, and capable of negotiating, executing, and personalizing tasks on behalf of individuals and organizations.

Being a cautious AI optimist, in INSTANT AI UPDATE 9, I wrote about AI hallucinations and the risks associated with AI providing “false but probable-sounding information.” This week’s headline about AI in med schools led me to research the growth of AI in medical education and identify the risks associated with this shift. 

Artificial Intelligence (AI) has rapidly evolved with the advent of large language models (LLMs). However, a critical and often misunderstood issue exists: “AI hallucinations”.  This phenomenon occurs when AI generates plausible but false or misleading information. As an AI author, I’ve experienced this many times. This has compelled me to adopt a “trust but verify” mantra.

This brief post provides an analysis of the three AI strategies. I hope this post offers clarity on how each platform positions itself in the competitive landscape and concludes with an evidence-based summary.

This post offers a detailed overview of Microsoft’s Copilot AI strategy, including 5 pros and 5 cons. Microsoft Copilot is an AI-powered assistant integrated into Windows and the 365 suite. Leveraging advanced AI technologies, including LLMs and Microsoft Graph, Copilot is designed to enhance productivity, automate repetitive tasks, and streamline workflows across applications, including Windows, Word, Excel, PowerPoint, Outlook, and Designer.

To our knowledge, this is the most comprehensive and up-to-date book on Apple’s AI strategies. As I’ve also shared, the book includes a Prologue with an overview and timeline of Apple Intelligence. It’s important to know that Apple has combined its ongoing work in machine learning (Core ML) with large-scale Artificial Intelligence under the Apple Intelligence umbrella in recent years. Read on to learn more...

This week, I've been working on the new INSTANT AI: APPLE INTELLIGENCE EDITION Chapter 8: Health & Fitness Insights Powered by AI. As part of this, I've been reviewing and considering the implications of Apple's latest information on Project Mulberry (and its successor project called Project Quartz). Here's my take on how this all comes together: 

INSTANT AI UPDATE 4: BUSY TIME FOR AI! Apple delays Siri, OpenAI expands tools, and six companies leading the AI charge. In this week's AI post, we delve into significant developments shaping the indust

I have long relied on basic accessibility tools that allow me to function normally in my academic and professional pursuits. I’ve been privileged to work with scores of differently-abled colleagues, including those with vision, hearing, emotional, physical, and intellectual challenges. As a consumer of accessibility capabilities and a longtime proponent of a genuinely equal-opportunity workplace, I believe that AI offers breakthrough capabilities for those of all abilities.