AI Blunder in the Courtroom
January 1, 2024
Good morning, and Happy New Year! We're thrilled to have 160,000 AI enthusiasts along for the ride in 2024. With robots brewing coffee and driving cars (allegedly), we think this year will promise even more incredible developments in AI than in 2023. Can't wait to break it down with you!
Thank you for being part of our community!
In today’s newsletter:
⚖️ Michael Cohen's AI Blunder
🚀 GitHub Launches Copilot Chat for All
🇨🇳 Nvidia’s China-Specific RTX 4090D
⚖️ Michael Cohen's AI Blunder: Mistaking Chatbot for Research Tool in Legal Filing
Michael Cohen, former lawyer for Donald Trump, mistakenly used Google’s Bard, an AI chatbot, for legal research, resulting in the citation of non-existent court cases in a legal document presented to a federal judge. Cohen, who sought to reduce his probation period after a guilty plea for tax evasion, inadvertently included these fake cases, believing Bard to be an advanced search engine. This error led to a federal judge questioning the legitimacy of the cited cases and potential sanctions for Cohen’s lawyer, David Schwartz. Cohen acknowledged his lack of awareness about the evolving legal technology and its risks. This incident highlights the growing influence of AI in legal proceedings, sometimes leading to unintended consequences like the misuse of AI-generated content.
This is becoming a persistent issue, and we’ve covered similar stories (see story four in that issue) multiple times here over the past few months. It will likely only worsen unless something is done about it…
🚀 GitHub Launches Copilot Chat for All: A New Era in AI-Assisted Coding
GitHub recently announced the general availability of Copilot Chat, a ChatGPT-like programming chatbot, now accessible to all users, including those in paid tiers and certain free categories like students and teachers. Built on OpenAI's GPT-4, Copilot Chat aids developers with real-time programming guidance, such as explaining concepts or writing tests. Despite its advancement, concerns remain over potential copyright issues and the absence of an opt-out mechanism for codebase owners. Additionally, while GitHub claims improvements in reducing AI-generated errors, there's still a need for vigilant human review. Competitively, GitHub's Copilot faces challenges in profitability and competition from rivals like Amazon's CodeWhisperer and various startups.
See Copilot chat in action 👇️
TOGETHER WITH VANTA
eBook: How to minimize third-party risk with vendor management
A robust vendor management program isn’t just required by compliance frameworks like SOC 2 and ISO 27001. It’s also a critical part of a holistic trust management strategy.
Implementing a vendor management program, however, has become more complex and challenging with the proliferation of SaaS tools and shadow IT. And many overstretched security teams are being asked to do more with less.
To stay compliant and secure — and deepen trust with customers and partners — security teams need a way to proactively manage vendor risk.
This guide from Vanta, the leading trust management platform, brings together perspectives from the frontlines of vendor security management. Get insights and best practices from security and compliance leaders.
🇨🇳 Nvidia’s China-Specific RTX 4090D: A Compliance Move Amid US Restrictions
Nvidia has introduced a new graphics card, the RTX 4090D, exclusively for the Chinese market to adhere to US export controls. This version is less powerful than the standard RTX 4090, featuring fewer CUDA cores (14,592 compared to 16,384) and a lower power draw (425W versus 450W). Despite these reductions, most other specifications remain similar between the two models. Nvidia’s move follows tightened U.S. restrictions on high-end chip exports to China, which also affect the company’s less powerful H800 and A800 AI GPUs. The RTX 4090D, slightly slower in gaming and creation tasks, is set to be available in China starting in January for approximately $1,836 USD. This release aims to meet the high demand for advanced graphics cards in China, where factories have reportedly repurposed the banned RTX 4090 for AI applications.
See the video below for a rundown on the new specs 👇️
LG has introduced a bipedal "AI agent" robot, designed as an all-around home manager and companion, set to debut at CES 2024. This innovative robot features remote home and pet monitoring capabilities, enhanced with voice and image recognition. It can greet users and play music tailored to their detected mood upon arrival. Read more here.
Google is considering a major workforce reduction, potentially affecting 30,000 employees, as part of integrating AI into its business processes, especially in the ad sales department. This shift towards AI could significantly reduce the need for human involvement in digital advertising and customer support services. This change is part of Google's broader strategy to realign its business in response to evolving technological landscapes despite the potential job cuts. Google CEO Sundar Pichai has acknowledged these challenges, emphasizing the need for adaptability in a transforming global environment. Read more here.
Director Christopher Nolan, known for his film 'Oppenheimer’ (among others), finds parallels between AI development and the nuclear era, cautioning against overestimating AI's capabilities. He emphasizes the need for accountability in AI usage, warning that extreme doomsday scenarios distract from current issues like copyright and employment impacts. Nolan's insights reflect broader concerns about regulating AI amid its rapid advancement in the tech industry. Read more here.
Companies are increasingly attracted to virtual influencers due to cost-effectiveness and control. A Financial Times report highlights Meta's ad analysis showing H&M's significant savings using virtual models, and an ad agency co-founder, Diana Nuñez, notes the high costs of real influencers. Virtual influencers also eliminate the challenges of demands, opinions, and potential controversies associated with real people. Read more here.
In China, once-promising robotaxi startups face challenges in achieving commercialization, leading them to seek alternative revenue streams. Hindered by safety, regulatory, and cost issues and influenced by the global geopolitical climate, these companies are shifting focus from fully autonomous taxis to more viable smart-driving solutions. They are also exploring partnerships with automotive manufacturers, government contracts, and international markets, as the viability of widespread robotaxi services remains uncertain. Read more here.
Starting 2024 with a win… finally!
We’re always trying to fake you out. Be careful out there!
Shall we continue? Oh, and come on… no googling, play by the rules 👀
Image 1
Which city is real? |
Image 2
Have cool resources or tools to share? Submit a tool or reach us by replying to this email (or DM us on Twitter).
*Indicates a sponsored link
The graph illustrates a predominant optimism among businesses regarding the role of big data analytics and AI in promoting job growth, with most respondents anticipating a positive impact from these technologies.
SPONSOR THIS NEWSLETTER
The Deep View is currently one of the world’s fastest-growing newsletters, adding thousands of AI enthusiasts a week to our incredible family of over 160,000! Our readers work at top companies like Apple, Meta, OpenAI, Google, Microsoft, and many more.
If you want to share your company or product with fellow AI enthusiasts before we’re fully booked, reach out below!
AI-Enhanced Proofreading
How It Works:
Upload Text: Provide a document or written content.
Proofreading Mode: Set the AI for grammar, style, or conciseness checks.
Receive Corrections: Get suggestions for improvements and apply as needed.
Why Use It?
Writing Quality: Enhance the clarity and readability of your content.
Language Polishing: Refine grammar and stylistic elements.
Productivity: Streamline the editing process for faster content turnaround.
Prompt Example:
"Proofread this document [insert document] for clarity and conciseness, focusing on correcting grammatical errors and refining the style to be more [insert specific style]."A meme before you go 👇
That's a wrap for now! We hope you enjoyed today’s newsletter :)
What did you think of today's email? |
We appreciate your continued support! We'll catch you in the next edition 👋
-Bot Buddy and The Deep View Team