Welcome to Daily Zaps, your regularly-scheduled dose of AI news ⚡️ 

Here’s what we got for ya today:

• 🤖 Apple two years behind in AI

• 😢 AI falsely accuses student of cheating

• 🖼️ Midjourney AI image edit feature

• 💸 Perplexity looking to fundraise at $8 billion valuation

Let’s get right into it!

Sign Up | Advertise | Tools

BIG TECH

Apple two years behind in AI

Some Apple employees believe the company is about two years behind in AI development compared to industry leaders. Apple introduced its AI feature set, Apple Intelligence, at WWDC24, including tools like an upgraded Siri and AI-driven notifications. Apple’s AI mostly runs on-device, requiring powerful chipsets like the A17 or M1, which limits the complexity of the models.

To bridge the gap, Apple partnered with OpenAI to integrate ChatGPT, which is 25% more accurate than Siri and can answer 30% more questions. Despite being behind, Apple is expected to catch up, with plans for Apple Intelligence to be available on all devices by 2026.

TECH

AI falsely accuses student of cheating

AI detection tools used by educators to identify AI-generated content are increasingly leading to false accusations of cheating, causing significant stress for students. For example, Moira Olmsted, a student juggling work, family, and studies, was falsely accused of using AI to complete her assignments due to her formulaic writing style, which mirrors AI patterns. Despite AI detectors being highly accurate, small error rates can have big consequences, particularly for neurodivergent students, non-native English speakers, and those with straightforward writing styles.

Tools like Turnitin, with a 4% false positive rate, can inadvertently penalize students. This growing reliance on AI detection in education is prompting concerns over its accuracy and fairness, with students resorting to documenting their writing process or using AI humanizers to avoid false flags.

STARTUPS

Midjourney AI image edit feature

Midjourney is set to release an upgraded web tool next week that allows users to edit any uploaded images using its generative AI, including retexturing objects and repainting colors based on captions. Initially, this tool will be available to a limited subset of the community with enhanced human and AI moderation to prevent misuse.

The release comes amid rising concerns over AI-edited images, deepfakes, and potential copyright infringement, as platforms struggle with how to label and control AI-modified content. While Midjourney has made some efforts to address these issues, including metadata embedding, it faces criticism for not fully embracing stronger image provenance tracking technologies like C2PA.

STARTUPS

Perplexity looking to fundraise at $8+ billion valuation

AI search engine Perplexity is reportedly in talks to raise $500 million at an $8+ billion valuation, more than doubling its valuation from $3 billion earlier this year after a funding round led by SoftBank. The company, which processes 15 million daily queries and generates $50 million in annualized revenue, uses AI in a chatbot-style interface to assist with web searches.

However, it has faced criticism from some news publishers for unauthorized web scraping, with The New York Times issuing a cease-and-desist letter. Despite these challenges, Perplexity aims to collaborate with publishers and continue growing amidst increased competition from AI-powered search products like OpenAI's SearchGPT.