118 | 👏 Everyone getting their own AI Models

In today’s Brainyacts:

1. Beyond boilerplate AI Use policies

2. Great interview with Microsoft and OpenAI CEOs

3. McKinsey releases its own AI model and other AI-related news

4. AI has created more images than all of history’s photographers combined and other news you can use

👋 to new subscribers!

To read previous editions, click here.

Lead Story

📑💻 Have you updated your website terms of service?

Companies are updating their terms of service to prevent web crawlers from using their data to train AI models because they want to protect their intellectual property and prevent others from using their data without permission. The New York Times updated its terms of services on August 3, 2023, to forbid the scraping of its content to train a machine learning or AI system. The updated TOS also prohibits website crawlers, which let pages get indexed for search results, from using content to train LLMs or AI systems.

In some cases, this content is replicated verbatim. Publishers, especially those with paywalls and healthy subscription businesses, are concerned that AI models will undermine their revenue streams by publishing repurposed content without credit and contribute to misinformation, degrading people’s trust in news.

It is not clear whether law firms should update their terms of service to prevent web crawlers from using their data to train AI models. What are some reasons why they might?

Law firms might consider updating their terms of service to prevent web crawlers from using their data to train AI models for several reasons. Here are some potential considerations:

1. Intellectual Property Protection:

   • Law firms produce a variety of content, including articles, blogs, whitepapers, and legal guides, that represent significant intellectual and creative investment. By updating their terms of service, they can assert control over this content and protect it from unauthorized use, similar to the concerns of publishers like The New York Times.

2. Maintaining Competitive Advantage:

   • Legal content, including insights and analyses by attorneys, is a key asset that differentiates one firm from another. Allowing competitors or other entities to train their AI models using this content could erode a firm’s competitive edge.

3. Preventing Misrepresentation:

   • AI models trained on a law firm’s data might inadvertently produce content that appears to be legal advice from that firm. This could result in misrepresentation of the firm's stance or opinions, which could potentially have legal consequences and damage the firm's reputation.

4. Data Privacy and Client Confidentiality:

   • Some law firm websites might contain sensitive information, such as case studies that, while anonymized, are based on real cases. Preventing web crawlers from using this data helps to add an additional layer of protection for client confidentiality.

5. Quality and Integrity of Legal Information:

   • Law firms have an interest in ensuring that legal information is accurate and used responsibly. If AI models are trained on their content and then produce incorrect or misleading information, this could have serious consequences for individuals or businesses that rely on this information.

6. Revenue and Monetization Concerns:

   • Some law firms monetize their content through subscription models or by using it as a value-added service for their clients. If AI models could repurpose and distribute this content freely, it could undermine these revenue streams.

7. SEO and Web Traffic:

   • If AI models repurpose content from a law firm's website and publish it elsewhere, this could potentially affect the search engine ranking of the original content. This is important as law firms often rely on their online presence to attract new clients.

8. Regulatory Compliance:

   • Law firms are subject to various regulations, including those related to advertising and client solicitation. Preventing unauthorized use of their content by AI models could help ensure that they remain in compliance with these regulations.

9. Control Over Use of Content:

   • Law firms may want to retain the ability to dictate how and where their content is used, and updating their terms of service to address AI and web crawling is a way of asserting this control more explicitly.

Now, I am not saying that firms should do this (many don’t have an easy-to-find terms of service page tbh) but I wanted to share these thoughts in case nobody has thought about this yet.

Spotlight Story

A 24 min video worth watching.

Recent interviews with both Microsoft CEO Satya Nadella and OpenAI CEO Sam Altman.

Touches on:

• It’s (ChatGPT) habit forming

• How each of them uses it in their own daily lives

• Concerns they have

• There are already lots of AI doing things we don’t know about, even before ChatGPT came about

• Can you actually turn AI off?

• Relationship between Microsoft and OpenAI

AI Model Notables

• End of ChatGPT Dominance? Enter Google’s Gemini

• OpenAI acquires start-up Global Illumination to work on core products

• OpenAI CEO says it does not train its models on anything using an API

• McKinsey unveils its own generative AI tool for employees: Lilli

• Google Chrome will summarize entire articles for you with built-in generative AI

• Arthur an open source tool to help companies find the best LLM for a job

• Allens launches enterprise version of ChatGPT, Airlie

News You Can Use:

➭ China’s new AI regulations begin to take effect

➭ Canadian government seeks input on voluntary code of practice for generative AI

➭ UK to Host AI Safety Summit in Bletchley Park in November

➭ Potential NYT lawsuit could force OpenAI to wipe ChatGPT to start over

➭ LinkedIn Says ChatGPT-Related Job Postings Have Ballooned 21-Fold Since November

➭ Thomson Reuters Officially Acquires Casetext for $650 Million as Deal Closes

➭ AI Has Already Created As Many Images As Photographers Have Taken in 150 Years

➭ Why OpenAI will not go bankrupt