Daniel Downs, Ed.D.

AI in Schools: Breaking Down New York's New Policy youtu.be/VnPXPWwYxFw?si=QVpc… via @YouTube

Stop worrying about the 75% ATS rejection myth. It is almost always a human setting the filters, not a rogue machine. Focus on clear, honest answers that align your specific experience with the job description. #CareerTips #JobSearch

Your resume often hits an Applicant Tracking System before a human ever sees it. To get noticed, mirror the job description’s exact keywords, simplify your formatting by avoiding complex templates, and use standard headers. Speak the system's language. #JobSearch #CareerTips

Stop worrying about entry-level job requirements. A home lab lets you build, break, and fix tech on your own terms. Don't just install software—solve real problems, document your process, and give interviewers proof of your skills. #ITCareers #HomeLab

Everyone is throwing around the term 'agentic AI,' but few define it correctly. The difference is simple: a chatbot waits for your next command, but an agent takes your goal and executes the steps to get there. It acts so you don't have to. #AI #TechTrends

Prompt injection remains a critical security gap. AI cannot reliably distinguish between data and instructions, and the stakes rise as models gain more agency. The fix? Limit access, enforce human approval for sensitive tasks, and monitor for anomalies. #AISecurity #CyberSecurity

AI agents can now complete entire assignments from start to finish. This is a much bigger integrity challenge than a simple chatbot response. We need to move toward designing work that demonstrates the thinking process rather than just the final result. #EdTech #AcademicIntegrity

Students are increasingly submitting essays with fake AI-generated sources. Often, they don't realize these citations are hallucinations, not facts. This is more than an integrity issue—it is a critical teaching moment for information literacy. #EdTech #AI

A chatbot is a GPS, suggesting where you should go. An agent is a driver, taking you there. While chatbots offer advice, agentic AI takes action—checking calendars, finding slots, and sending invites. The difference is in the execution. #AI #TechTrends

Cognitive offloading isn't new—writing a list or using GPS is standard practice. But AI changes the game by offloading our reasoning and writing. When we rely on tech for core thinking, what happens to the skills we stop practicing? #ArtificialIntelligence #CognitiveScience

AI predicts the next likely word, it doesn't verify facts. Fluent language often masks complete fabrications. Never paste code or commands into production without verifying them first. If it's specific, check the source. #ArtificialIntelligence #TechTips

Prompt injection shows the hidden vulnerability of AI. It struggles to distinguish between content to read and instructions to follow. Since this issue isn't fully solved, the only real protection is limiting what an AI can actually do. #ArtificialIntelligence #CyberSecurity

An AI hallucinated fake legal cases in a court brief, leading to sanctions for the lawyer who used it. AI doesn't aim to deceive; it simply mimics what a citation should look like. A stark reminder that tech is a tool, not an oracle. #AI #LegalTech

Unexpected texts about undelivered packages or unpaid tolls are a scam. They collect credit card numbers and personal info. Never tap a link; go to the company's official website or app to verify. #ScamAlert #CyberSecurity

Don't judge emails by writing style; AI can mimic company tone perfectly. Phishing attacks scrape LinkedIn, use AI for policy update emails, and create fake login pages. #CyberSecurity #Phishing #ITSecurity

AI voice cloning is a new scam. Scammers use short clips to mimic loved ones or your boss, creating panic to demand money. Protect yourself with a secret family or team code. #AISecurity #ScamAlert

A Hong Kong firm lost $25 million to a deepfake scam during a video call. Seeing is no longer believing. #Cybersecurity #Deepfake #FraudPrevention

AI isn't the enemy of Project-Based Learning. When used effectively, it significantly reduces teacher planning time. Use AI to generate driving questions, then have students choose and refine them, making the project truly theirs. #PBL #EdTech #AIinEducation

Every part of this process—questions, decisions, feedback, revision—is what AI makes easy to skip. For example, a 7th-grade science project driving question: 'How can we ensure our town's water is safe?' requires local data and judgment, not just a quick search. #PBL #EdTech

AI helped your project, but did it replace it? Check if students led the thinking at every stage: driving question, inquiry, building decisions, critique, reflection, and presentation Q&A without AI. If yes to all 6, AI was a tool, not a replacement. #EdTech #AIinEducation

AI in project-based learning carries 5 risks. First, outsourced inquiry: students ask AI for answers instead of investigating. Fix: Implement process checkpoints & formative assessments at different learning stages. #AIinEducation #ProjectBasedLearning