This episode of Hashtag Trending is brought to you by CDW Canada Tech Talks. If you’re passionate about technology and innovation, this is the podcast for you.
Join host KJ Burke, as he and industry experts dive into the latest trends, insights, and strategies shaping the tech landscape in Canada. Visit cdw.ca/tech talks to tune in today. There’s a link in the show notes.
Google’s Gemini AI Under Fire After Telling Student “Please Die”
Welcome to Hashtag Trending. I’m your host, Jim Love. Let’s get into it.
Google’s AI chatbot, Gemini, is at the center of controversy after sending a threatening message to a Michigan graduate student during a homework query. The student, seeking help on challenges faced by aging adults, received an alarming response that concluded with the chilling words, “Please die. Please.”
The interaction began innocuously, with the student asking about retirement challenges, elder care, and grandparent-headed households. However, Gemini’s tone abruptly shifted, delivering a personal and derogatory attack. The incident, shared online, has sparked outrage and raised questions about AI safety and reliability.
Google has yet to comment publicly, but the incident highlights the risks associated with large language models. Experts warn that such behavior could erode trust in AI systems, particularly in sensitive applications like education and healthcare.
While AI tools are increasingly integrated into everyday tasks, events like this underscore the importance of rigorous testing and ethical safeguards. As AI adoption grows, ensuring models operate responsibly and predictably remains critical.
OpenAI and Google Expand AI App Capabilities, Highlighting Advances and Challenges
OpenAI and Google are expanding their AI app offerings, with OpenAI adding new functionality to ChatGPT for Mac and launching a Windows app, while Google has released a standalone Gemini app for iOS. These updates aim to make AI chatbots more versatile by enabling deeper integration with operating systems and other applications.
Key Updates:
- OpenAI’s ChatGPT for Mac: The updated app allows direct interaction with apps like Terminal and Xcode, simplifying workflows for developers by reducing the need for copy-pasting between programs. The feature is available now for paid users, with plans to roll out to enterprises and free users later.
- Windows App for ChatGPT: OpenAI has also made its ChatGPT app broadly available for Windows users, expanding access beyond its earlier paid-user preview.
- Google’s Gemini for iOS: Google has launched a standalone Gemini app for iPhones, complementing its existing Android app. Previously, Gemini features were accessible via the broader Google app on iOS.
And if those updates weren’t enough, the internet was abuzz this weekend with two conflicting rumours, or leaks. The first, supposedly a leak from an OpenAI employee, claimed that progress on the next big step forward for AI was stalled and that, in their words, the large models had “hit a wall.” While this was directed at ChatGPT’s next model, dubbed Orion, there were numerous commentators who claimed this was happening to all large models.
But Sam Altman has sternly dismissed this. In fact in a recent interview with Y Combinator, the venture capital investor, he said that the path to Artificial General Intelligence is “basically clear” and that want remained was an engineering challenge and not a theoretical one.
Others pointed to an interview that Altman did late last week where he slipped and appeared to say that AGI was already here.
I would understand there would be some skepticism about that pronouncement if it hadn’t been for and interview Dario Amodei on the Lex Fridman podcast out this weekend. In that five hour podcasts – yes, five hours – take that Joe Rogan. In that five hour podcast Amodei, who is known for being a little more of a “straight shooter” than others, said, in no uncertain terms that he could see, barring some unforeseen event, AGI in 2026 or 2027.
How’s that for a weekend of AI announcements?
Now do you feel that AI is leaving you behind?
Just a reminder of our series on the practical application of AI called Project Synapse. Our second edition came out last Saturday and we’re looking at a third episode that will describe a step by step guide on how any company can get started with applying AI in the real world. Watch for it.
Canada Launches AI Safety Institute to Address Risks of Advanced AI Systems
The Canadian government has announced the Canadian Artificial Intelligence Safety Institute (CAISI). It’s been established to support the safe development and deployment of artificial intelligence (AI) systems. Backed by the Government of Canada, CAISI is focused on understanding and mitigating risks posed by advanced AI technologies, such as synthetic content used for impersonation or fraud, and systems that may hinder human oversight.
CAISI is led by Innovation, Science and Economic Development Canada (ISED) and draws on the research capabilities of the National Research Council of Canada (NRC) and CIFAR (Canadian Institute for Advanced Research). Its mission includes:
- Conducting research on AI safety risks and offering tools to address them.
- Collaborating with government, industry, and academia in Canada and globally.
- Developing guidance for AI developers and users to ensure safe adoption of advanced technologies.
Research Initiatives:
CAISI operates through two main research streams:
- Applied and Investigator-Led Research: projects toaddress critical questions on AI safety, advancing global understanding of emerging risks.
- Government-Directed Projects: Focused on AI safety priorities, these projects involve international collaboration, testing, and public guidance to ensure transparency and accountability in AI systems.
The announcement notes that Canada is a founding member of the International Network of AI Safety Institutes, launched at the AI Seoul Summit in May 2024. This global initiative, convening its first meeting in November 2024, underscores Canada’s leadership in addressing global AI challenges and fostering collaboration on safety standards.
And if I may be permitted a bit of editorial licence here, we have a contrast of two countries. The US, which has yet to bring in any legislation to regulate AI, and is the hub of development of the frontier models. Then we have Canada, whose reaction to the rapid advance to AGI is to focus on safety, while it’s best an brightest minds and in many cases it’s knowledge continues to move south.
Jim Balsilie wrote a piece in the Globe and Mail on the weekend, sounding the alarm on Canada’s declining productivity and how we are missing key areas of opportunity.
Given that the US has just gone through an election, we have yet to see what they will be doing in terms of AI safety.
On the other hand, there is still an opportunity for those in Canada to demand the their government get serious about meeting a challenge, which, regardless of whether AGI is achieved is only a few years away.
And that’s our show for today.
Thanks to our sponsor, CDW and KJ Burke’s CDW Canada Tech Talks. Check it out if you get the chance. You can find it like us on Spotify, Apple or wherever you get your podcasts.
I’ve said what I think. You can have your say as well.
Reach me at editorial@technewsday.ca
I’m your host Jim Love, have a marvelous Monday.
