236 | 👓 🚨 AI glasses that show your home address by seeing your face

In today’s Brainyacts:

1. Not another GenAI disclosure requirement!

2. OpenAI’s Canvas demo

3. AI glasses that uses faces to get addresses! and other AI model news

4. Chat with your 60 year old future self and more news you can use

   👋 to all subscribers!

To read previous editions, click here.

Lead Memo

🤦 📑 Court blunders with a GenAI disclosure form

Generative AI is seeping into our daily workflows with many unlikely to notice but starting October 21, 2024, if you’re filing something in the Butler County, Pennsylvania Court of Common Pleas, you’ll have to include an affidavit about whether you used generative AI in preparing your documents.

On the surface, this might seem like a step toward maintaining integrity in legal filings, but let’s be real: this is a bureaucratic hoop-jump that feels disconnected from the realities of modern legal practice.

The singling out of generative AI is not only unnecessary—it’s a bit ridiculous when you think about how integrated AI has become in the tools most lawyers use every day.

Here’s the deal: Lawyers and pro se litigants have always been responsible for verifying the accuracy of the filings they submit. Whether you’re using a case book, a Westlaw search, or even Google, the obligation to make sure your case law and citations are correct has never changed. But now, with this affidavit, the court is asking you to disclose if you used AI in your preparation, and verify that all citations have been checked by a human. My question is—what’s the point? Shouldn’t we already be verifying that our work is accurate, AI or no AI?

Take a moment to consider the average legal workflow today. If you’re doing legal research, chances are you’re using Westlaw or LexisNexis, both of which now have AI baked into their platforms. It’s not always obvious, and frankly, most of us don’t stop to think, “Hey, is this result AI-generated?” We just assume the platform is doing its job and focus on reviewing the cases or statutes it pulls up. Same goes for Google, which as of late 2024 is sneaking generative AI into its search results. So, whether we like it or not, AI is part of our daily workflow.

What’s the court really asking for here? Are they expecting us to track every AI interaction we’ve had in the process of drafting a motion or brief? How would you even know if AI was involved in a Lexis search or a Westlaw headnote summary?

The affidavit is asking us to disclose something that, in many cases, we can’t reasonably track or may not even be aware of. This demonstrates a key disconnect and gap in understanding. It shows the court needs to learn more about generative AI before it slaps a silly requirement on us.

Here’s the rub: This new affidavit doesn’t prohibit the use of AI—it just adds an extra layer of paperwork to verify that, if we did use it, we personally checked our sources. It’s like the court wants to appear vigilant about AI, but what they’ve really created is a rubber stamp that doesn’t address any real concerns.

I get it. There’s the lingering fear about AI “hallucinating” cases or misrepresenting legal authorities. But affidavit isn’t going to stop lazy prompting or sloppy research, and it won’t make people suddenly better at fact-checking. What it will do is give us more to sign, more to attach, and more reasons to roll our eyes.

Look, I understand what the court is trying to do here—they want to make sure AI isn’t being used recklessly in legal filings. That’s fair. But this isn’t the way to go about it. AI is just another tool in the toolkit, and like any tool, its usefulness depends on the person using it. The focus should be on the substance of the filing, not the method. If I’m filing a motion, what matters is whether I’ve checked the law and the facts, not whether I used AI to help me write the first draft.

Spotlight

🔥🤯 OpenAI just changed how we interact with AI

OpenAI introduced a new way to interact with ChatGPT on Thursday: an interface it calls “canvas.” The product opens a separate window, beside the normal chat window, with a workspace for writing and coding projects. Users can generate writing or code directly in the canvas, then highlight sections of the work to have the model edit. Canvas is rolling out in beta to ChatGPT Plus and Teams users on Thursday, and Enterprise and Edu users next week.

• In tests, using GPT-4o with Canvas led to a 30% accuracy and 16% quality boost compared to using the model without the interface.

In this video, I walk you through OpenAI's new Canvas feature, available for Plus and Teams subscribers. It’s a fresh way of interacting with GPT-4, and here are the highlights:

1. Inline Editing: You can highlight and edit sections of text, ask GPT-4 to rewrite or add details, and format directly within the interface.

2. Interactive Sidebar: Ask follow-up questions and get instant feedback on specific sections without leaving the interface.

3. Customization: Adjust reading level, change text length, and polish the final output in real-time.

4. Formatting Tools: Add emojis, bold text, and finalize your document without needing external tools.

Try it out and see how it improves your workflow!

AI Model Notables

► Google just announced the introduction of ads to its AI Overview search summaries. "We’ve been carefully testing ads in AI Overviews for relevant queries. We’ve seen that people are finding ads directly within AI Overviews helpful because they can quickly connect with relevant businesses, products and services to take the next step."

► The AI glasses that reveal anyone’s personal details—home address, name, phone number, and more—just from looking at them.

► OpenAI raises $6.6B in the largest VC round ever, reaching a post-money valuation of $157B. Microsoft, NVIDIA, and SoftBank were among the participating investors.

► Microsoft Copilot Voice mode: The rollout for Voice Mode began earlier this week. If you are interested in accessing it, create a Microsoft account, download the free Copilot app if you plan on using it on your phone, or update the app if you already have it downloaded. 

► Microsoft’s mammoth AI bet will lead to over $100 billion in data center leases.

► Meta drops new text-to-video tool: Meta Movie Gen.

News You Can Use:

➭ Army testing robot dogs armed with AI-enabled rifles in Middle East.

➭ Generative AI and Information Warfare (webinar): A case study focused on how the Chinese military may adopt generative AI to interfere in Taiwan's democracy.

➭ Podcast: AI Companies are opting you in by default - using your data to train their models and trying to opt out isn’t always clear and simple.

➭ AI for Legal Aid: How to supercharge legal services organizations: 2 case studies.

➭ Future You: A conversation with an AI-generated future self reduces anxiety, negative emotions, and increases future self-continuity.

➭ The Cancer AI Alliance formed a $40M collaboration between major medical institutions and tech giants like Microsoft, AWS, Nvidia, and Deloitte to advance AI-driven cancer care.

➭ Judge dismisses antitrust claims against Thomson Reuters in legal search battle.

Who is the author, Josh Kubicki?

Some of you know me. Others do not. Here is a short intro. I am a lawyer, entrepreneur, and teacher. I have transformed legal practices and built multi-million dollar businesses. Not a theorist, I am an applied researcher and former Chief Strategy Officer, recognized by Fast Company and Bloomberg Law for my unique work. Through this newsletter, I offer you pragmatic insights into leveraging AI to inform and improve your daily life in legal services.