Generative AI: A Jack of All Trades?
More than one and a half years after OpenAI released ChatGPT, starting a craze around large language models and generative AI, Apple has officially boarded the AI hype train. On Monday, the iPhone maker unveiled a whole suite of AI features that will be baked into the upcoming versions of iOS, iPadOS and macOS. Apple of course wouldn’t be Apple if it didn’t have its own take on the matter, so it didn’t just call those features AI features but a “personal intelligence system” called Apple Intelligence.
With a focus on privacy, as most tasks will be completed on-device, Apple Intelligence showcases the wide-ranging capabilities of AI, as it will help users craft texts, create images, take actions across app and make everyday tasks easier by understanding context and drawing appropriate conclusions.
The wide range of new features Apple announced on Monday highlights not only the many strengths of AI at its current stage, but also one weakness. While AI is a jack of all trades, doing a lot of things quite well, a true killer feature has yet to emerge, perhaps explaining the limited usage of ChatGPT and other AI tools so far. While it’s great to play around with these novel tools, few users really rely on AI in their everyday routines, which of course, Apple’s entrance into the field could change.
According to a recent survey conducted by YouGov on behalf of the Reuters Institute for the Study of Journalism, “playing around or experimenting” is in fact one of the most popular use cases for generative AI tools so far and even so, just 11 percent of more than 12,000 adults surveyed across six countries have used an AI tool for that purpose. Answering factual questions is another widespread use case, although “widespread” is a bit of a stretch at 11 percent. Summing up several activities, 28 percent of respondents had used an AI tool to create media in some way, while 24 percent had used artificial intelligence to get information. With writing tools, an ”Image Playground” and deep Siri integration, Apple is addressing both these use cases with Apple Intelligence – but only time will tell what will emerge as the killer feature.
Dataset
| Category | Respondents who have tried a generative Al tool for the following applications (in %) | |
|---|---|---|
| Playing around or experimenting | Creating Media | 11% |
| Answering factual questions | Getting Information | 11% |
| Asking advice | Getting Information | 10% |
| Writing an email or letter | Creating Media | 9% |
| Making an image | Creating Media | 9% |
| Generating ideas | Getting Information | 9% |
| Playing around or experimenting | Getting Information | 9% |
| Writing an essay or report | Creating Media | 8% |
| Summarizing text | Getting Information | 8% |
| Creative writing | Creating Media | 7% |
| Seeking support | Getting Information | 7% |
| Recommendations | Getting Information | 6% |
| Translations | Getting Information | 6% |
| A job application/interview | Creating Media | 5% |
| Programming or coding | Creating Media | 5% |
| Making a video | Creating Media | 4% |
Data sources
12,217 adults surveyed in Argentina, Denmark, France, Japan, the UK and the US; Mar. -Apr. 2024
