New Bill to Address Cyber Security and AI Systems at Public Sector Entities
The provincial government introduced Bill 194, the Strengthening Cyber Security and Building Trust in the Public Sector Act, 2024, which would lay the foundation to regulate the use of AI in public institutions. Though much of the details are left to future regulations, the Bill sets the stage for what will be a full set of rules governing public institutions' use of AI, including restrictions, responsibilities to protect and safeguard personal information, and individuals' rights when their information has been compromised.
What is AI? OECD Updated Definition of Artificial Intelligence
AI just isn't what it used to be. In November 2023, The OECD released an updated definition of Artificial Intelligence that has helped provide a clearer understanding of what should and should not be considered AI. At the core of the new definition is an emphasis on adaptability and autonomy.
While this doesn't provide a clear set of AI versus non-AI buckets, it does mean that the test for tools that may have traditionally been defined as AI is clearer. In the past, we may have assumed that the ability of an Optical Character Recognition tool to turn a scan of a fax into a searchable document was leveraging AI, but that would not necessarily be the case. If it isn't able to learn and adapt from each scan, then it wouldn't fall into the new definition. That said, developers can certainly evolve their tools from non-AI into the new definition of AI. So, what was once considered AI can reclaim that label in the future.
Google's Gemini Goes Pro, ChatGPT Turns the Big 4o, and Claude comes to Canada
The last few weeks have been full of developments with Google, OpenAI and Anthropic all making important announcements about the functionality and accessibility of their LLMs.
ChatGPT 4o
OpenAI launched the next iteration of the ChatGPT model on May 13, naming it ChatGPT 4o. The "o" stands for omni. Why omni? This new model is able to process multiple modalities without having to first convert to text. This is a significant advancement in the way we can interact with AI, making it much faster at responding to our requests. At this time, Chat GPT 4o is publicly limited to processing text and audio inputs. But, at the launch they highlighted the ability to analyze video as well. View OpenAI's launch.
Google Gemini
The day after OpenAI launched ChatGPT 4o, Google made a number of upgrades to their Gemini Model, with their new Gemini 1.5 Pro model taking centre stage. Gemini 1.5 Pro, which requires a monthly subscription, has dramatically increased the amount information it can process at one time. It is able to analyze up to 2 million tokens, which is roughly equivalent to a 1,500-page PDF. By comparison, OpenAI's newest model, ChatGPT 4o can process about 128,000 tokens and Anthropic's Claude has a 200,000 token context window. Google also showed off, but did not widely release, a number of features that are coming to their Gemini platform, including the ability to verbally interact with text uploaded to Gemini through their Bookworm tool, a new custom GPT competitor called Gems, as well as Veo for generating high-definition AI videos. Watch a summary of the Google Gemini announcements.
Claude Comes to Canada
Anthropic, the AI company that has built their brand around AI safety and transparency, has finally made their Claude model available in Canada. For those looking for an alternative to ChatGPT and Google Gemini, Anthropic's Claude may be of interest. As with other AI platforms, Claude offers free and paid models. Paid models should be explored if you are using the platform for supporting tasks in your legal practice. Some protections and customization are limited to paid customers.
Learn more about the impact of generative AI on the legal landscape - specifically how lawyers are making use of Microsoft Copilot in their practices - with a dedicated OBA series coming up this fall. Stay tuned to the Real Intelligence on AI hub and this newsletter for details!
Real Intelligence on AI Program Spotlight
AI Trial Advocacy: Your Comprehensive Guide to Handling the Evidentiary Implications of Artificial Intelligence
September 27 and November 1, 2024
With the expanding capabilities and use of artificial intelligence (AI) both in legal practice and beyond, it is critical that litigators are well equipped to navigate the evidentiary issues that arise when AI is involved. From AI-generated evidence to experts' reliance on AI in formulating their opinions to unpacking and challenging underlying AI technologies, the implications of the use of AI and the strategies you need to respond are varied. Ensure you have the insights and expertise you need with this highly practical two-day program.
Sign up to be the first to know more.
AI Trial Advocacy: Your Comprehensive Guide to Handling the Evidentiary Implications of Artificial Intelligence
September 27 and November 1, 2024
With the expanding capabilities and use of artificial intelligence (AI) both in legal practice and beyond, it is critical that litigators are well equipped to navigate the evidentiary issues that arise when AI is involved. From AI-generated evidence to experts' reliance on AI in formulating their opinions to unpacking and challenging underlying AI technologies, the implications of the use of AI and the strategies you need to respond are varied. Ensure you have the insights and expertise you need with this highly practical two-day program.
Sign up to be the first to know more.
Chat GPT Series for Lawyers: Now On Demand
Did you miss any installments of our Chat GPT Series for Lawyers? Access them on-demand by clicking program title and video tab from the Series page.
Latest Analysis from Lawyers
OBA members have been sharing their AI expertise and opinions in an array of intriguing Section articles. We've rounded up some of the most recent AI reading:
- The Dual Edge of AI: Promise and Perils for Cyber Defence by Allessia Chiappetta and M. Imtiaz Karamat
- Prompt Attention Pays Off with Generative AI tips from Monica Goyal and Al Hounsell
- Unmasking Deception: Canada Weighs in on the Dark Side of Deepfakes by Jennifer R. Davidson and Victoria Di Felice
… and From Around the Web
From the curious to the serious, this AI news is making headlines:
- If A.I. Can do Your Job, Maybe it Can Also Replace Your C.E.O. (The New York Times)
- Lenovo, Mastercard join government's voluntary AI code of conduct (Financial Post)
- A Safety Check for Open AI (The New York Times)
- Senior judge predicts AI will transform the role of experts (The Law Society Gazette)
- Google research shows the fast rise of AI-generated misinformation (CBC.ca)
OBA Names New Innovator-in-Residence
The OBA has announced Colin Lachance as our 2024-25 Innovator-in-Residence. In this role, he will spearhead a groundbreaking initiative aimed at assisting lawyers with effectively integrating AI into their practice to better serve their clients.
AI Robot Speaks at Commencement
University World News: "D'Youville University in the United States embraced the future of technology by having an AI robot named Sophia speak at its commencement ceremony."