Windows Copilot Newsletter #16 - Copilot for Finance, Claude 3, and AI worms...
Windows Copilot Newsletter #16
Microsoft lures CFOs; Claude 3 bests GPT-4; and the arrival of the AI worms...
G'day and welcome to the sixteenth edition of the Windows Copilot Newsletter, where we curate all of the most interesting stories from the emerging field of AI chatbots. This has been a very big week, so let's dive right in...
Top News
Copilot for Finance: In its drive to 'Copilot all the things', Microsoft announced Copilot for Finance: "Copilot for Finance draws on essential context from your existing financial data sources, including traditional Enterprise Resource Planning (ERP) systems, such as Microsoft Dynamics 365 and SAP, and the Microsoft Graph..." If it works as promised, this could become a must-have planning tool for CFOs - just as Excel was a generation ago.
Anthropic launches Claude 3: The 'other child' of the 2020 OpenAI divorce, Anthropic released the third version of their Claude language model. In its top-of-the-line version, Claude 3 'Opus' claims to best even GPT-4-Turbo in standard benchmarks. We've been waiting for something to come along and knock the crown off OpenAI - who better than the forgotten child? Read Anthropic's announcement here.
Oh no you didn't: OpenAI fired back at now-estranged-cofounder Elon Musk's lawsuit claiming that the firm has departed from its initial vision of 'AI for everyone', providing some choice quotes detailing how the world's second richest man may have twisted a few facts to suit his argument. Read their post (with Elon's damning emails) here.
Is Copilot Designer dangerous? An engineer at Microsoft made a report to the US Federal Trade Commission claiming that Microsoft's 'Copilot Designer' image generation tool can generate very unsafe images. Read about that here.
Top Tips
How to disable Copilot: Lifehacker provides a step-by-step guide for those folks who may not want their Windows desktops slowly and completely colonised by Microsoft's AI tentacles.
How to use ChatGPT, Copilot and Gemini: For those new to this whole area, Axios provides a useful how-to. Read that here.
Safely and Wisely
Here come the AI worms: A 'worm' is a networked computer virus, capable of migrating from machine to machine, spreading its infection as it travels. Researchers shared with WIRED their own research constructing prompts that create an "AI worm" - infecting a chatbot that can then infect other chatbots. This could be a serious security threat for all AI chatbots - read about it here.
CloudFlare for your chatbot: Cloud cybersecurity provider CloudFlare introduced a new product designed to protect AI chatbots from denial-of-service attacks. CloudFlare's "Firewall for AI" will also inspect the responses generated by a chatbot - ensuring they're not leaking any confidential information. Read about it here.
Longreads
How do the large language models that underlie AI chatbots work? Truth be told, no one really knows for sure. It's a topic of active research, as reported in this fascinating article from MIT's Technology Review.
What are the prospects for AI chatbots in an educational setting? This detailed whitepaper from Educause Review looks at the opportunities, challenges - and necessary policy settings.
Understanding the Microsoft Copilot Pro Value Proposition
Last month, Drew Smith and I released our first Wisely AI white paper - designed to help organisations evaluate whether the features in Copilot Pro justify handing Microsoft a hefty AUD $45 per month per person subscription fee. Read or download the white paper here.
We're hard at work on our next white paper - all about how organisations can de-risk their use of AI. More on that next week - when we’ll be back with more AI chatbot news!
If you learned something useful in this newsletter, please consider forwarding it to someone else who might benefit.
Thanks!
Mark Pesce
www.markpesce.com // Need help with AI?