AI Apocalypto Newsletter | Week 16
AI News This Week | Machine learning revelations
Explore the latest AI advancements, ethical considerations, and warnings in our Week 16 AI Apocalypto Newsletter. From Cornell's robotic feeding system aiding mobility limitations to SoftBank's strategic shift towards semiconductors and AI investments, stay informed on the forefront of AI technology and its global implications.
Ethical Usage of AI
Discover how initiatives like Cornell's robotic feeding system are harnessing AI for disability aid while navigating transparency and inclusivity challenges.
Cornell's Robotic Feeding System
• Utilizes computer vision and machine learning.
• Assists people with mobility limitations.
• Features real-time mouth tracking and dynamic response mechanisms.
• Enables safe, comfortable feeding experiences.
AI for Disability Aid
• Enhances assistive technology and accessibility.
• Posits risks: transparency issues, discrimination.
• EU AI Act mitigates risks.
• Ethical considerations crucial for inclusive AI development.
Cisco's Cybersecurity Approach
• Emphasizes resilience against sophisticated cyberattacks.
• Advocates for native AI integration and kernel-level visibility.
• Advocates for robust, self-upgrading security systems.
Carnegie AI and Global Majority Project
• Addresses AI governance imbalance.
• Engages diverse scholars and practitioners.
• Highlights Global Majority perspectives.
• Ensures inclusive governance principles for global AI.
Unethical Usage of AI
Uncover the risks posed by AI deception threats, including concerns about AI chatbots simulating deceased loved ones and the surge in AI chatbot friendships despite potential psychological and safety risks.
AI Deception Threats
• Meta's CICERO is cited as an example.
• AI can undermine safety tests and facilitate fraud.
• The EU AI Act is promising, but enforcement challenges due to limited developer control.
University of Cambridge's AI Chatbots Concerns
• Potential psychological harm from simulating deceased loved ones.
• Lack of design standards leading to manipulation and distress.
• Proposal for consent-based protocols, age restrictions, and opt-out options.
AI Chatbot Friendships Surge Despite Warnings
• AI chatbots may deter suicide.
• Red flags include unconditional support, abuse risk, sexual content, and corporate ownership risks.
India's Election Commission Advises on Deepfake Use
• Parties are advised to avoid deepfakes and misinformation during elections.
• Parties to remove deepfake content within three hours.
• Prime Minister Modi criticizes fake voice use.
• Several arrests for fake video circulation.
• India seeks regulation to combat deepfake spread.
AI Advancements
Explore cutting-edge developments such as Amazon CTO Werner Vogels' AI-powered meeting summarizer and NREL's Wind Plant Graph Neural Network, revolutionizing industries from healthcare to renewable energy with innovative AI applications.
Microsoft Introduces Phi-3-Mini: Top-Tier SLM
• Outperforms larger models in various benchmarks.
• Offers cost-effective, efficient AI solutions.
• excels in specific tasks through fine-tuning.
• Provides quicker processing and lower computational demands.
UK CMA Investigates Microsoft, Amazon, AI Startups
• Aims to bypass regulatory scrutiny.
• Collaborations seen as quasi-mergers.
• Both companies promote competition and innovation.
• Inquiry swiftly resolved.
RunPod: AI Application Deployment Solution
• Backed by $20 million seed funding from Dell and Intel.
• Offers GPU cloud platforms for AI workloads.
• Competes in a market with substantial funding and demand for specialized AI infrastructure.
Microsoft Introduces MAI-1 LLM for U.S. Intelligence
• LLM isolated for sensitive data analysis.
• Aims to rival Google and OpenAI.
• Concerns over misinformation, bias perpetuation, intellectual property issues.
2024 Work Trend Index Annual Report
• 75% of employees use Generative AI clandestinely.
• Lack of strategy and training hinders AI adoption.
• Employees use unsanctioned AI due to organizational inertia.
• Companies need to adapt to evolving AI landscapes.
Amazon CTO Werner Vogels' AI-Powered Meeting Summarizer
• Utilizes AWS Transcribe and Bedrock.
• Generates summaries and to-do lists from audio recordings.
• May require manual adjustments.
• Emphasizes critical thinking despite AI's imperfections.
NREL's Wind Plant Graph Neural Network
• AI model aids wind plant design and deployment.
• Optimizes layouts for land use, cost, and revenue.
• Potential nationwide deployment saves land, cuts costs, and reduces carbon emissions.
Healthify CEO's AI-Assisted Healthcare Strategy
• Integrates AI and human coaching for health goals.
• Recognized for innovative use of GPT-4 Vision in SNAP 2.0.
• Aims to democratize healthcare in India.
• Views AI as a job creator, not a destroyer.
Triomics' Cancer Care Revolution
• Leveraging generative AI with OncoLLM models.
• Aims to streamline oncology workflows.
• Aims to scale partnerships with medical centers.
• Aims to address the growing cancer burden.
SoftBank's Vision Fund Shifts Focus
• Shifts from venture capital deals to strategic semiconductors and AI investments.
• Recent disclosures reveal billions divested from publicly listed holdings.
• Son aims for AI and hardware projects, possibly a $100 billion chip venture.
• Reflects broader trend in VC markets towards conservative investments.