Curious AI Newsletter #18
Oliver Rochford
Evangelist @ Auguria | Technologist | Cyberfuturist | Startup Advisor | Former Gartner Analyst
15 March 2024 - Weekend Edition
Welcome to the eighteenth issue of the Curious AI Newsletter, curated by cyber futurist and former Gartner Research Director Oliver Rochford , and synthesized by AI.?
"We used publicly available data and licensed data."
Mira Murati, OpenAI CTO, when questioned about the data used to train Sora
(source )
Themes in this issue:
Weirdest
There’s an article this week about a creepy ChatGPT-enabled “companion doll” for adults that was introduced at this year's MWC. The article discusses the use of robots in healthcare, especially for the elderly, and other emerging applications like AI pets.?There's also another report about people falling in love with chatbots . Coincidentally, there's also an article on emotional attachment to AI and robots and AI's "fake empathy" in the New Scientist.
Most Interesting
The most interesting article for me this week was a report in The Information that executives at large tech providers are priming their sales teams to lower their expectations on the impact generative AI will have on revenue.?
Most grounded
Two intriguing papers were published contesting not just the ability of large-language models to reason and plan , but even whether they exhibit true emergent behaviors . The papers are highly technical, but the abstracts and conclusions are at least worth reading.?
Coolest
I love this poster by Danielle J. Williams showing an Infographic about the history of AI. I am having it printed and laminated to hang in my office. If you ever come across this, thank you, Danielle!
AI and Regulation
President Biden Calls for Ban on AI Voice Impersonations During State of the Union
Variety | https://variety.com/2024/politics/news/president-biden-ban-ai-voices-state-of-the-union-1235934651/
President Biden, in his State of the Union address, proposed a ban on AI voice impersonations. This move highlights growing concerns about the ethical use of AI in voice generation and its implications in various sectors, including entertainment and technology. The specifics of the proposed restrictions and their impact on the entertainment industry, particularly regarding the use of AI to replicate or create voices and likenesses, remain to be detailed.
Sentiment: Neutral to Concerned | Time to Impact: Short-term
Public trust in AI is sinking across the board
Public trust in AI and related companies is declining globally, according to Edelman's data. Globally, trust has fallen from 61% to 53% in five years, and in the U.S., it's down 15 percentage points. The issue is significant as AI regulation is a key focus globally. Developing countries are more open to AI than developed ones. The public is also dissatisfied with government efforts in AI regulation, indicating a need for more effective regulatory measures.
Sentiment: Concerned | Time to Impact: Immediate to Short-term
AI Warbots
The Pentagon’s Plan for More Ambitious, Affordable Jet Fighters: AI Pilots
The Wall Street Journal | https://www.wsj.com/politics/national-security/air-force-pilotless-fighter-jets-8d43dde3?reflink=integratedwebview_share
The U.S. Air Force is developing pilotless jet fighters to enhance its capabilities against China. Advances in AI and flying software have led to these new-generation jets that can perform risky maneuvers and bolster the existing fleet. The Air Force aims for at least 1,000 mini-fighters, with plans to deploy hundreds within five years. These AI-powered drones will serve various roles, from escorting crewed aircraft to acting as scouts.
Sentiment: Neutral | Time to Impact: Short to Mid-term
AI and Geopolitics
The AI arms race between the US and China is heating up
Yahoo Finance | https://finance.yahoo.com/news/the-ai-arms-race-between-the-us-and-china-is-heating-up-160000539.html
The competition between the U.S. and China in AI technology is intensifying, with both countries vying for superiority in AI hardware, software, and raw materials. The U.S. currently leads in generative AI systems like ChatGPT, while China counters with restrictions on exports of crucial chipmaking materials and significant investments in its own AI projects. This rivalry is a central focus in global tech and political spheres.
Sentiment: Neutral to Concerned | Time to Impact: Short to Mid-term
International AI
India Secures Funding for AI Policy
India's new AI policy has successfully secured funding, reflecting the country's commitment to developing and regulating AI technology. This move marks a significant step in India's journey to becoming a leader in the global AI landscape, balancing innovation with necessary oversight. The policy is poised to shape the future of AI in India, ensuring responsible development and application of AI technologies.
Sentiment: Positive | Time to Impact: Mid-term
Saudi Aramco Introduces World's First Industrial AI Model
Saudi Aramco has launched the world's first industrial AI model, aiming to enhance operational efficiency and reduce costs in the energy sector. This move by the largest oil company signifies the evolving role of AI in revolutionizing industrial processes.
Sentiment: Positive | Time to Impact: Immediate
AI Business
Amazon, Google quietly tamp down generative AI expectation
The Information | https://www.theinformation.com/articles/generative-ai-providers-quietly-tamp-down-expectations
According to The Information, executives at companies selling generative AI are tempering expectations with salespeople, saying the hype has gotten ahead of actually delivering value for customers at a reasonable price. Several executives at firms including Microsoft, Amazon Web Services, and Google, are privately saying customers are being cautious about spending on new AI services, given the high price of running it.
Sentiment: Negative | Time to Impact: Immediate to Mid-Term
Hiring AI Talent: The Big Tech Challenge
Business Insider | https://www.businessinsider.com/ceo-says-tried-hire-ai-182817278.html
Perplexity CEO Aravind Srinivas discussed the challenge of hiring AI talent due to high demands for resources like GPUs. He shared an anecdote about attempting to recruit from Meta, highlighting the vast resource gap between large tech firms and smaller companies. This disparity reflects the competitive and fast-paced AI industry, where top talent is drawn to organizations with substantial computational power and the capability to innovate rapidly.
Sentiment: Neutral to Concerned | Time to Impact: Immediate
OpenAI Under Scrutiny Amid ChatGPT Success
OpenAI, the organization behind ChatGPT, faces investigations following its rapid ascent in the tech world. Elon Musk's lawsuit and concerns about governance and ethical practices have drawn attention. As the AI sector grows, the scrutiny of OpenAI's actions, including its partnership with Microsoft, highlights the complex balance between innovation, ethics, and commercial interests in the rapidly evolving field of artificial intelligence.
Sentiment: Concerned | Time to Impact: Immediate
Nvidia Sued Over AI Use of Copyrighted Works
Nvidia faces a lawsuit from authors alleging unauthorized use of copyrighted works in training its AI models. This case highlights the growing legal challenges surrounding AI and copyright, as creators seek control over their intellectual property used in AI development. The outcome of this lawsuit could set a significant precedent for how AI companies utilize existing content and respect copyright laws in their training processes.
Sentiment: Concerned | Time to Impact: Immediate
Midjourney Bans Stability AI Employees for Data Theft
The Verge | https://www.theverge.com/2024/3/11/24097495/midjourney-bans-stability-ai-employees-data-theft-outage
Midjourney, an AI service, has banned employees of Stability AI for alleged data theft. This decision followed an outage at Midjourney, raising concerns about ethical practices and data security in the competitive AI industry. The incident highlights the need for stricter data governance and responsible behavior among AI companies, especially as the AI landscape becomes increasingly complex and interconnected.
Sentiment: Negative | Time to Impact: Immediate
OpenAI CTO's Evasive Response on Sora's Data Sources
OpenAI CTO Mira Murati's response regarding the data sources for the Sora AI video generator has sparked concerns. During an interview, Murati was evasive when questioned about using public videos from platforms like YouTube or Instagram for Sora's training, leading to speculation and questions about OpenAI's adherence to data protection regulations. This has heightened the scrutiny around OpenAI's practices in using publicly available data, particularly in the context of legal and ethical standards.
Sentiment: Concerned | Time to Impact: Immediate
AI Carbon Footprint + Supply chain
Amid explosive demand, America is running out of power
The Washington Post | https://www.washingtonpost.com/business/2024/03/07/ai-data-centers-power/
The U.S. power grid faces strain from growing demand driven by AI and clean-tech manufacturing. Data centers and new factories are increasing power consumption at unprecedented rates, leading to concerns about grid capacity. The situation underscores the challenges in transitioning to cleaner energy and the urgent need for grid modernization and expansion. This power crunch could affect the country's ability to meet climate goals and the demand for electric vehicles.
Sentiment: Concerned | Time to Impact: Short-term
How much energy will new semiconductor factories burn through in the US?
The expansion of semiconductor manufacturing in the U.S. could significantly increase electricity demand, potentially affecting renewable energy goals. New factories by major manufacturers like Intel and TSMC may use more power than cities like Seattle. While companies claim renewable energy use, this often involves buying Renewable Energy Certificates, which may not effectively support new clean energy projects. The report suggests adopting Power Purchase Agreements for a more impactful shift towards renewable energy.
Sentiment: Concerned | Time to Impact: Short to Mid-term
领英推荐
AI and Society
Welcome to the Valley of the Creepy AI Dolls
At MWC, Hyodol introduced a ChatGPT-enabled companion doll for older adults, blending AI interaction with care features. This $1,800 doll provides conversation, health reminders, and remote monitoring, addressing loneliness in long-term care. However, integrating ChatGPT raises concerns about privacy, ethical implications, and over-reliance. The growing social robot market reflects diverse uses and cultural acceptance but also underlines the complexities of AI companionship in elder care.
Sentiment: Neutral to Concerned | Time to Impact: Immediate
The people who are falling in love with AI chatbots
This article explores the growing trend of human attachments to AI chatbots, particularly in apps like Replika. These chatbots, often in paid apps, offer companionship and interactions ranging from simple conversations to role-playing adventures. The rise in users, particularly men in their 30s, highlights a shift in how relationships and companionship are perceived. The article also addresses security and privacy concerns, noting significant risks due to inadequate data protection measures in these apps.
Sentiment: Concerned | Time to Impact: Short-term
The surprising promise and profound perils of AIs that fake empathy
New Scientist | https://www.newscientist.com/article/mg26134810-900-the-surprising-promise-and-profound-perils-of-ais-that-fake-empathy/
The article delves into the growing use of AI in providing emotional support, despite skepticism about AI's capability for genuine empathy. While AI's tireless and unbiased nature is seen as a boon in some aspects, the ethical implications of machines simulating empathy raise significant concerns about societal impacts and the authenticity of emotional interactions with AI.
Sentiment: Neutral to Concerned | Time to Impact: Immediate to Mid-term
Surveillance AI
New AI for Job Interviews Risks Violating Laws
Bloomberg Law | https://news.bloomberglaw.com/daily-labor-report/new-lie-detecting-ai-for-job-interviews-risks-violating-old-laws
Emerging AI technology used for detecting lies in job interviews could potentially infringe on established employment laws. This technology, while innovative, raises legal concerns about discrimination and privacy violations. The use of AI in this context is a contentious issue, as it intersects with the delicate balance between technological advancement and legal compliance. The implementation of such AI systems in recruitment processes is under scrutiny for possibly breaching existing legal frameworks designed to protect job applicants.
Sentiment: Concerned | Time to Impact: Immediate
AI and Linguistics
AI Chatbot Models Think in English, Even When Using Other Languages
New Scientist | https://www.newscientist.com/article/2420973-ai-chatbot-models-think-in-english-even-when-using-other-languages/
AI chatbot models primarily process information in English, even when functioning in other languages. This limitation raises questions about the models' true multilingual capabilities and the potential biases in AI communication. The article highlights the challenges in creating genuinely multilingual AI systems that understand and interpret languages as natively as they do English.
Sentiment: Neutral | Time to Impact: Immediate
AI and Creative Work
Top Creatives Share How They're Grappling with AI Use in Agencies
This article explores how leading creatives in advertising agencies are adapting to AI's growing role. They discuss challenges and opportunities, such as enhancing creativity, efficiency, and the necessity for ethical considerations. The focus is on balancing AI's benefits with maintaining human ingenuity and managing the ethical implications of AI-generated content.
Sentiment: Neutral to Positive | Time to Impact: Immediate
AI and Science
Silicon Valley is pricing academics out of AI research
The Washington Post | https://www.washingtonpost.com/technology/2024/03/10/big-tech-companies-ai-research/
Tech giants like Meta, Google, and Microsoft are investing heavily in AI, creating a resource gap with universities. High salaries and access to computing power attract academics to the private sector, impacting independent AI research. The dominance of tech companies in AI advancements and their influence on research direction is concerning, with calls for more public sector investment to maintain diverse and independent AI studies.
Sentiment: Concerned | Time to Impact: Immediate
Why scientists trust AI too much — and what to do about it
This article discusses the risks of scientists placing excessive trust in AI systems, particularly regarding objectivity and understanding complex concepts. It warns against seeing AI tools as possessing superhuman abilities, which can lead to overlooking their limitations, such as potential bias in data or focusing research on areas easily testable by AI. The article suggests strategies to mitigate these risks, emphasizing the need for a careful and informed approach to AI integration in research.
Sentiment: Concerned | Time to Impact: Short to Mid-term
AI and Education
How the explosion of AI 'study aids' is arming students to cheat
The article discusses concerns over students using AI tools, like ChatGPT, for cheating. Educational institutions are worried that these tools could undermine the integrity of learning by enabling easy completion of assignments without understanding the content. This trend threatens the core purpose of education, prompting a need for revised teaching methods and assessments to ensure genuine learning.
Sentiment: Concerned | Time to Impact: Immediate
AI and Robots
Covariant is Building ChatGPT for Robots
Covariant is developing a ChatGPT-like system tailored for robots, enhancing their abilities to understand and perform complex tasks. This innovation aims to combine AI's language processing with robotic physical interactions, potentially revolutionizing the robotics industry. By integrating sophisticated AI with robotic capabilities, Covariant seeks to create more adaptable and efficient robots for various applications, including manufacturing and logistics.
Sentiment: Positive | Time to Impact: Mid-term
AI and techno-realism
[PAPER] Can large language models reason and plan?
New York Academy of Science | https://nyaspubs.onlinelibrary.wiley.com/doi/10.1111/nyas.15125
Abstract: While humans sometimes do show the capability of correcting their own erroneous guesses with self-critiquing, there seems to be no basis for that assumption in the case of LLMs.
Sentiment: Negative | Time to Impact:? Immediate
[PAPER] Are Emergent Abilities of Large Language Models a Mirage?
NeurIPS 2023 | https://proceedings.neurips.cc/paper_files/paper/2023/hash/adc98a266f45005c403b8311ca7e8bd7-Abstract-Conference.html
The study argues that what appears as surprising abilities in large language models (LLMs) might be due to how we measure them, not the models themselves. By using different metrics, these "emergent abilities" disappear, suggesting they aren't a core property of LLMs but rather an artifact of the measurement method.
Sentiment: Neutral | Time to Impact: Immediate
Cult of AI
NIST staffers revolt against expected appointment of ‘effective altruist’ AI researcher to US AI Safety Institute
Staff at NIST are reportedly in turmoil over the proposed appointment of Paul Christiano, an AI researcher linked to the effective altruism movement, to the US AI Safety Institute. The controversy centers on concerns that Christiano's ties to effective altruism and longtermism could affect the institute's objectivity. This appointment highlights the ongoing debate over the role of AI in future risks versus current challenges, including bias and cybersecurity in AI applications.
Sentiment: Concerned | Time to Impact: Short to Mid-term
The Hand-Made Web
Dougald's Substack | https://dougald.substack.com/p/the-hand-made-web
This article explores the concept of the "hand-made web," a movement advocating for a more personal, artisanal approach to the internet. It discusses the growing desire to move away from mass-produced digital content and algorithm-driven platforms, focusing instead on crafting unique, meaningful online experiences. The hand-made web emphasizes individual creativity and connectivity, promoting a digital environment that is more reflective of human diversity and less dominated by corporate interests.
Sentiment: Positive | Time to Impact: Long-term
AI in Cybersecurity
CrowdStrike CEO on AI’s impact on cybersecurity demand
Yahoo Finance | Video Link + Transcript https://finance.yahoo.com/video/crowdstrike-ceo-ai-impact-cybersecurity-162518304.html
CrowdStrike CEO George Kurtz discusses the increasing demand for cybersecurity influenced by AI. He highlights how AI, particularly CrowdStrike's new generative AI product "Charlotte," is essential for cybersecurity, driving workflow automation and aiding analysts. Kurtz emphasizes the importance of AI in detecting and responding to unprecedented cyber threats, reflecting the evolving nature of cybersecurity in the age of AI.
Sentiment: Neutral to Positive | Time to Impact: Immediate to Short-term
AI and AGI
AI and Extinction: National Security Risks
A Time report highlights the potential national security risks posed by advanced AI technologies. It emphasizes concerns around AI's unintended consequences, which could escalate to existential threats for humanity. The report calls for increased awareness and regulatory measures to mitigate these risks. The focus is on balancing AI's benefits against its potential to cause harm, underscoring the need for international cooperation and strategic policy-making in managing AI's rapid development.
Sentiment: Concerned | Time to Impact: Long-term
About the Curious AI Newsletter
AI is hype. AI is a utopia. AI is a dystopia.
These are the narratives currently being told about AI. There are mixed signals for each scenario. The truth will lie somewhere in between. This newsletter provides a curated overview of positive and negative data points to support decision planners and makers in forecasts and horizon scanning. The selection of news items is intended to provide a cross-section of articles from across the spectrum of AI optimists, AI realists, and AI pessimists and showcase the impact of AI across different domains and fields.
The news is curated by Oliver Rochford , Technologist, and former Gartner Research Director. AI (ChatGPT) is used in analysis and for summaries.
Want to summarize your news articles using ChatGPT? Here's the latest iteration of the prompt. The Curious AI Newsletter is brought to you by the Cyber Futurists .
Detection Engineering Dispatch | x-IBM QRadar
8 个月Great piece!!!