The week in AI: Sports Illustrated CEO fired over AI scandal & Google's Gemini woes

Presented by

Welcome to The Dispatch! We are the newsletter that keeps you informed about AI. Each Thursday, we aggregate the major developments in artificial intelligence; we pass along the news, useful resources, tools or services, and exciting projects in open source. Even if you aren’t an engineer, we’ll keep you in touch with what’s going on in AI.

"At the bottom [of the page] there would be a photo of a person and some fake description of them like, 'oh, John lives in Houston, Texas. He loves yard games and hanging out with his dog, Sam.' Stuff like that," they continued. "It's just crazy."

The AI authors' writing often sounds like it was written by an alien; one ‘Ortiz’ article, for instance, warns that volleyball "can be a little tricky to get into, especially without an actual ball to practice with."

SI additionally terminated operations president and chief operating officer Andrew Kraft, media president Rob Barrett and corporate counsel Julie Fenster in wake of the gaffe.

Google is facing backlash after admitting its flashy Gemini demo video was heavily edited, misleading consumers with the viral video. The slick demo wowed viewers with Gemini's capability in recognizing visual cues and interacting vocally with a person in real time. In reality, frames were extracted from footage, and then researchers entered written prompts for Gemini off-camera, massaging the interactions.

While some are defending brief demo embellishment as common practice, critics argue it misrepresents capabilities. It was a rollercoaster week for Google and Gemini - the demo quickly went from being viewed as a landmark achievement to what now looks to be a bit like smoke and mirrors.

Perhaps most embarrassingly of all for Google, content creators are now replicating the video using ChatGPT in real-time - prompting without edits to highlight how far ahead GPT-4V is even when compared to a staged Gemini demo.

The glasses communicate responses verbally and save photos and AI responses in the Meta View phone app. Meta's Chief Technology Officer, Andrew Bosworth, envisions a future where the glasses will have low-power sensors to trigger AI awareness automatically. These glasses are some of the first steps towards more advanced, multimodal, wearable AI technology - along with pins like Humane’s Ai pin.

European negotiators have reached further agreement on the EU’s sweeping AI Act. The Act, first drafted before ChatGPT existed back in 2021, has faced major revisions during negotiations. Key issues include the regulation of foundation models (general-purpose AIs, including language models) and the use of AI in law enforcement, particularly concerning facial recognition.

European firms, particularly in France, have expressed concerns that the regulations might still hinder innovation. Additionally, the laws could influence American AI companies' operations in the EU, making them more cautious in releasing cutting-edge products. It raises questions about investment in European AI ventures and the movement of AI talent.

You can read more details about the agreements from the European Parliament website.

Another controversial use of artificial intelligence has popped up in politics: Pennsylvania Democrat Shamaine Daniels is utilizing "Ashley," an AI-powered political campaign caller, in her congressional campaign against Trump-aligned Republican Representative Scott Perry. Ashley is capable of conducting personalized conversations with many voters simultaneously. Over a weekend, Ashley called thousands of Pennsylvania voters, tailoring conversations to their key issues and communicating in over 20 languages.

While this type of AI tool provides underdog candidates like Daniels with new outreach capabilities, it raises concerns about potential disinformation in politics . Civox, who developed Ashley, is aware of the ethical implications and decided to give Ashley a robotic voice in addition to openly disclosing her AI nature.

More in AI this week:

• OpenAI and Axel Springer (POLITICO, Business Insider, etc.) are partnering to bring global news content to ChatGPT

• The reality of AI - vanishing white-collar jobs

• 3 reasons why the future of AI relies on women

• Why Google’s Bard could succeed or fail - and what that means for the future of AI chatbots

• Pro-China YouTube network used AI to malign U.S., report finds

• As ChatGPT gets “lazy,” people test “winter break hypothesis” as the cause

• Microsoft is going nuclear to power its AI ambitions

• Ex-OpenAI board member Helen Toner breaks silence on Sam Altman drama

eBook: How to minimize third-party risk with vendor management

A robust vendor management program isn’t just required by compliance frameworks like SOC 2 and ISO 27001. It’s also a critical part of a holistic trust management strategy.

Implementing a vendor management program, however, has become more complex and challenging with the proliferation of SaaS tools and shadow IT. And many overstretched security teams are being asked to do more with less.

To stay compliant and secure — and deepen trust with customers and partners — security teams need a way to proactively manage vendor risk.

This guide from Vanta, the leading trust management platform, brings together perspectives from the frontlines of vendor security management. Get insights and best practices from security and compliance leaders. 

Thanks for reading our newsletter today - see you next Thursday!