AI features are all of the rage these days. You can’t go a couple of seconds without hearing about a new startup or Chinese intuitive such as Deepspeak disrupting the entire space. E-Readers have hardly any AI features, but the e-notebook space is starting to embrace it in a big way. Maybe they hope to capitalize on a trend, without accounting for the user experience. When buying a new digital paper products, does AI drive consumer interest, or does it even matter?
Bigme has been making waves in the phone and e-notebook space by going all in on whatever AI buzz is going around. A number of their products have Deepspeak R1, ChatGPT, and numerous others. They say its to write code, document generating or solving math problems. One of the downfalls is that none of this can be uninstalled and people worry about their privacy.
Likely, the biggest company jumping on the AI bandwagon is Amazon, with the Kindle Scribe. They have implemented a new notebook summation feature, powered by generative AI, allows you to quickly create a synopsis of your notes so you can bring your ideas off the page and turn them into actions. If you’re scribbling in a hurry, or don’t have the tidiest handwriting, the new AI-powered refinement feature will convert your handwritten notes into text. All of these features are sent to the Amazon severs can sometimes take as a few minutes for anything to happen.
The most egregious form of AI has to be Onyx Boox. According to a report Azure GPT-3 on Boox devices got silently replaced a model by Bytedance, the company behind TikTok. It is a glorified chat bot that will give official state sponsored messages by China or her allies. It won’t say anything negative, about even North Korea. The outcry over Boox’s AI assistant has ebbed after Boox reportedly switched back to OpenAI’s GPT-3 via Microsoft Azure, according to another user’s post in the Boox subreddit.
As you can see, AI features can be a mixed bag, but it doesn’t have to be this way.
The Good AI
There is a whole host of AI that could benefit users and instead of marketing AI like its a godsend, it could benefit users. I would love to see personalized reading recommendations, such as genre-based suggestions, employing sophisticated algorithms to analyze a user’s reading history, preferences, and reading speed to curate highly relevant book suggestions.
A-driven recommendation engines could even become more nuanced. For instance, if a user frequently reads historical fiction and has recently shown interest in World War II topics, the AI might recommend newly released novels or critically acclaimed books within that sub-genre.
This capability goes beyond basic collaborative filtering and delves into content-based analysis, understanding themes and styles within books themselves. This new level of predictive capability means less time spent searching for the next read and more time immersed in a captivating book perfectly matched to individual preferences.
When it comes right down to it, book recommendations are the weakest part of reading. Amazon, B&N, Kobo, Google or Apple Books spend millions of dollars on an editorial staff, making book recommendations. Sure, you might see books by the same author or books you might want to read. Until we get good AI making proper recommendations it is really hard to know what to read next.
I have always found that the best way to implement AI, is to fool the user that no AI is present.
Michael Kozlowski is the editor-in-chief at Good e-Reader and has written about audiobooks and e-readers for the past fifteen years. Newspapers and websites such as the CBC, CNET, Engadget, Huffington Post and the New York Times have picked up his articles. He Lives in Vancouver, British Columbia, Canada.