Windows Copilot Newsletter #22 - Meta goes all in; Microsoft invests; ChatGPT hallucinates
Windows Copilot Newsletter #22
Meta releases LLaMA 3 and chatbots ‘all the things’; Microsoft makes a big investment with geopolitical implications; ChatGPT continues to confabulate scientific citations…
G’day and welcome to the twenty-second Windows Copilot Newsletter, where we curate all the most important news from the rapidly expanding-field of AI chatbots. It’s been a huge week of news, so let’s dive right in…
Top News
Meta goes all in on AI: This week, Meta released LLaMA 3, its latest-and-greatest language model, and simultaneously announced integrations of ‘Meta AI’ into its entire product line - Facebook, Messenger, WhatsApp, even a dedicated website at meta.ai. Has ChatGPT finally met a worthy competitor? Have the ‘big four’ become the ‘big five’? Read the story here.
Microsoft makes a strategic AI investment: The New York Times reported that Microsoft had made a USD $1.5 billion investment in UAE AI startup G42, in a deal reportedly brokered by the Biden Administration - to keep the technology from falling into Chinese hands. The AI Cold War just keeps on going.
Gemini in Chrome: Just like Microsoft, Google wants all to ‘Gemini all the things’. Its latest upgrade provides direct access to Gemini from the Google Chrome address bar. Just put ‘@gemini’ in the front of the bar, and type your prompt. Too easy - but will users want this sort of accessibility? Read that story here.
Copilot on Windows 10: Microsoft’s massive rollout and upgrade of Copilot to the billion-or-so Windows 10 desktops has well and truly begun. It will be some months before it’s completed - and Microsoft will have reached a billion and a half Copilot-connected computers.
Special focus on AI and the Law
Looking AI in the Eye: Will lawyers adopt AI in their practice? It’s early days, and as this article from the Boston College Law School Magazine reports, some lawyers are enthusiastic, others less so.
One in two lawyers use AI: Another article from the Australian Financial Review profiles lawyers who integrate AI into their practice - and how they do it.
Top Tips
Handy tips for Copilot in Windows 11: A gentle guide from Gizchina.
5 Copilot Pro Features: If you’re just getting started with Copilot Pro, this guide will be a big help.
Safely and Wisely
ChatGPT hallucinates scientific citations: According to PsyPost, false citation rates generated by ChatGPT for psychology research are ‘staggering’ - anywhere from 6% to 60%. Why you’d ever use a chatbot for citations is the question we should be asking.
The NSA wants you to secure your AI: The USA’s premiere agency for cryptography and signals analysis, the National Security Agency released a comprehensive guide on how to use AI securely. Grab it here.
Longreads
A.I. has a Measurement Problem: Which chatbot is smartest? Most reliable? Least likely to hallucinate? So many possible metrics - and so little agreement. The New York Times investigates.
The Rise of the Chief A.I. Officer: A brand new role for a new class of business tools. The Financial Times explains.
‘De-Risking AI’ white paper
Wisely AI is pleased to announce the release of our second white paper. ‘De-Risking AI’ looks at five of the risk organisations confront when using AI tools: anthropomorphising AI chatbots; malicious and commercially sensitive training data; hallucinations; privacy, data security and data sovereignty; and prompt attacks. It’s a must-read for any organisation working with or thinking about working with AI tools. Read it or download it here.
(photo Drew Smith)
More next week - we’ll be back with the latest AI chatbot news!
If you learned something useful in this newsletter, please consider forwarding it along to a colleague.
Mark Pesce
mark@safelyandwisely.ai // Wisely AI