AI's Trough of Disillusionment

Will AI Pay Off?

Author

AI For All
August 02, 2024

In partnership with

Hello readers,

Welcome to another edition of This Week in the Future! Generative AI might be entering the trough of disillusionment as adopters and investors grow concerned about ROI. How did we get here and is there a way out? Let’s find out!

The Trough of Disillusionment

Made with DALL·E 3

According to research from Gartner, 30% of generative AI projects will be abandoned after the proof of concept stage by the end of 2025. Businesses are struggling to justify the costs associated with deploying generative AI, which range anywhere from $5 million to $20 million, and find it difficult to translate productivity enhancements into financial benefit. Other reasons cited include poor data quality and inadequate risk controls.

A study conducted by Upwork found that generative AI is not even delivering on said productivity enhancements. 47% of employees surveyed say they have no idea how to achieve the productivity gains their employers expect, and 77% say generative AI actually decreases their productivity. Upwork blames organizations for having “outdated models and systems” that can’t see the true glory of generative AI.

IBM published a guide to generative AI for CEOs that acknowledged the difficulties of realizing gen AI value, declaring that “there’s no such thing as an all-purpose gen AI model.” The guide attempts to guide one to the promised land. Meanwhile, good ol’ industrial AI is doing just fine, with adoption and new use cases soaring.

Adoption is one half of the story. How are AI providers faring? Meta reported better-than-expected results in its Q2 earnings call despite its massive gen AI investment. This is due to the strength of its ad business. Nevertheless, Meta made it clear that they don’t expect gen AI to be a meaningful driver of revenue.

Our Take

What does all this mean? We may be entering the trough of disillusionment for generative AI, where experiments and implementations fail to deliver. There are two questions that need to be asked: why is this happening and will things recover?

Two theories are typically proposed for why. The first is that businesses don’t know how to properly implement gen AI. The second is that gen AI is just not that good, which is what at least one CIO concluded when cancelling Microsoft Copilot. As for recovery, it will depend entirely on model improvement. For now, AI is best for augmenting specific tasks, not overhauling an entire enterprise. Focus on the value you get out of it.

🔥 Rapid Fire

An entirely new way to present ideas

Gamma’s AI creates beautiful presentations, websites, and more. No design or coding skills required. Try it free today.

📖 What We’re Reading

“Machine translation (MT) has come a long way. The field has seen remarkable advancements from the early rule-based systems to the advent of neural networks. However, despite all the progress, traditional MT models still face significant challenges. They often struggle to understand context, handle complex language structures, or adapt to different domains.”

Source: AI For All

“Like various business sectors across the modern digital landscape, contact center operators are grappling with the opportunities posed by Artificial Intelligence (AI), which is playing a pivotal role in reshaping their operations. Driven to swiftly capitalize on AI’s potential, operators are actively exploring immediate and long-term benefits.”

Source: AI For All

💻️ AI Tools and Platforms

  • Telow → AI for business intelligence

  • o9 Solutions → AI-powered integrated planning

  • Grasp → The world’s first AI assistant for finance

  • Leya → AI-powered workspace for legal

  • Granola → AI notepad for meetings