In the years to come, 2023 may be remembered as a year of generative AI hype, with ChatGPT becoming a widely adopted technology and high expectations for AI-powered advancements. However, in 2024, there will be a need to recalibrate these expectations due to setbacks and limitations of generative AI. Evidence will emerge that these AI models can provide false information and are prone to hallucination. The notion of quick fixes to these problems will prove optimistic, as the architecture of these models makes it difficult to anchor predictions to known truths. The focus will shift to understanding which human tasks can be augmented by generative AI and the additional training needed by workers. Furthermore, the use of generative AI in social media and online search will lead to increased manipulation, misinformation, and mental health problems. While there will be more AI startups and open source initiatives, the dominance of tech giants like Google and Microsoft/OpenAI will continue, leading to disappointment in many applications based on these models. Calls for antitrust and regulation will intensify, but meaningful regulation is not expected to arrive in 2024.
| Signal | Change | 10y horizon | Driving force |
|---|---|---|---|
| Generative AI hype | Shift from unrealistic expectations to recalibration | More evidence of false information and hallucinations | Hype and lack of understanding |
| Disappointing productivity improvements | Recognition of the need for human tasks augmentation | Focus on implementation and additional training for workers | Faulty implementation of generative AI |
| Recognition of the limitations of predicting words for complex human cognition | Shift in understanding of AI capabilities | More realistic expectations of intelligence development | Recognition of limitations in AI architecture |
| Increased manipulation and misinformation online | Growing manipulation and misinformation | More time spent on screens and mental health problems | Monetization through individualized digital ads |
| Emergence of duopoly in the AI industry | Domination by Google and Microsoft/OpenAI | Reliance on gargantuan models and disappointment of apps | Lack of competition and disappointing performance |
| Calls for antitrust and regulation | Intensified calls for regulation | Lack of progress in antitrust action, discussions around new laws and regulations | Concerns about technology and bipartisan discussions |