CES 2025: AI Disappointments and 'IRL AI Slop'

The Pervasive, Yet Questionable, Application of AI in 2025

Despite advancements, a clear understanding of practical AI applications remains elusive for many companies, as evidenced by observations at this year’s Consumer Electronics Show (CES).

The event showcased AI-integrated products – including kitchen appliances and baby cribs – where the necessity of artificial intelligence was debatable.

Examples of Dubious AI Integration

Consider Spicerr, a spice dispenser featuring a touchscreen interface. This “intelligent” device aims to learn user preferences and suggest recipes.

However, Spicerr’s practicality is questionable, as it lacks grinding capabilities and relies on expensive, non-refillable proprietary capsules. The fundamental demand for an AI-powered salt and pepper shaker is also unclear.

Dreo’s ChefMaker 2, an AI-powered air fryer, was another demonstration at the show.

This concept, while less far-fetched than Spicerr, offers features like recipe extraction from cookbooks via page scanning and automated calculation of cooking times and temperatures.

Yet, it’s uncertain whether the air fryer-buying public actively requested cookbook scanning functionality. Many users haven't considered this feature a necessity.

CES also presented even more unusual AI-driven products.

The Case of Razer’s Project Ava

Razer’s Project Ava, named after a robotic character from the film “Ex Machina,” functions as an “AI gaming copilot.” It attempts to assist players without directly controlling the game.

Ava captures screenshots and provides guidance, such as suggesting when to dodge attacks. However, it has drawn criticism for being trained on gaming guides without acknowledging the original authors.

Furthermore, the system introduces a noticeable delay and interrupts game audio with its instructions, proving to be a distraction.

The core question remains: is there a genuine demand for such a product, and will users adopt it regularly, or be willing to pay for it?

Industry Hype and Limitations

The proliferation of these unconventional AI products at CES suggests an industry-wide overestimation of the technology’s potential. U.S. AI companies secured $97 billion in funding last year alone.

Vendors are experimenting with various AI applications, driven by the potential for significant returns and minimal risk.

These efforts often encounter the inherent limitations of current AI technology. Determining viable use cases remains a significant challenge, frequently leading to inflated promises and underwhelming results.

Existing AI models, like ChatGPT, are prone to errors, image generators can produce historically inaccurate depictions, and AI-generated videos often exhibit visual inconsistencies.

Consequently, we are left with a landscape of impractical AI integrations – air fryers, spice dispensers, and gaming copilots – that don’t necessarily align with consumer needs, but are achievable with relatively limited research and development investment.

Optimistically, future iterations will yield more meaningful applications.

Stay informed with TechCrunch’s AI newsletter! Subscribe here to receive it weekly in your inbox.