ChatGPT meets iOS

“Apple Nears Deal With OpenAI to Put ChatGPT on iPhone”, 彭博资讯 says.

A deal that could have a huge influence over the whole tech world. There are still no official comments from the both sides, so we can only guess and discuss about the main functions and features of this convergence. So, we all have a room for creativity.

As it seems, whilst design is almost solid and doesn’t change a lot, our mobile devices are becoming more and more complex from the hard- and software prospective. The high-quality photoshoots in the darkness or predictive text typing are the smallest parts of modern development, that aren’t surprising us anymore. But both of them and other features can be enhanced by the #AI. How? As long as I’m not the AI-expert or engineer, I can only imagine the results of this infusion, e.g. Siri becomes smarter and smarter.

United together millions of iPhones with integrated #ChatGPT will create a distributed #supercomputer that can solve the most difficult calculations in real time. The scale will differ from the Chat usage because of the number of smartphone users that is bigger than the PC-users. As mentioned in Pew Research Center Report (December 6, 2022), the 18-country median of 93% represents a vast majority of adults in the advanced economies that use the internet (people that report using the internet or owning a smartphone).

?

According to the Pew Research Center Fact Sheet (January 31, 2024), an average number of the U.S. adults owning a smartphone is about 90%. We can’t extrapolate this number to the other countries, but we can assume a tendency of the demographic influence on the smartphones and internet markets.

The higher number of smartphone users is, the higher are the capacities available to the ChatGPT in iPhones, that can be used for calculations or another complex measurements and researches. By the way, as we see from the fact sheet, majority of the different educating schools’ students use the smartphones, that can give to AI more scientifical information for self-learning from the students’ educating programs and information based on the students’ social interactions (that could be used both for behavioral predictions and for anthropology researches). Astonishing field for researches.

I guess, AI could also fix bugs in our devices, so we’ll reduce our actions with the service stations, while the main technical problems and bugs are scored in the log-lists with the predicted scripts and solutions. As long as these processes (problem-repeat, reason, solution) are thousands of times analyzed and fixed in the instructions, they can be automated and implemented in the iOS as the self-repair and maintaining feature. That will also reduce #Apple costs on technical support leaving the engineers solving the most difficult problems occurring with the devices’ physical conditions.

By collecting our data, our phone better know and understand us, hence more accurate marketing offers and recommendations are provided. The AI could also work with our photos, videos and texts to create better diaries or give a “third-party” opinion on our interactions. Maybe Siri even could offer us some self-prepared text-answers based on our personal texting styles and we’d need just to approve the messages? Siri gets instrument to talk instead of us with our contacts using our patterns – the deepfake concept expands enormously, and even when we’re dead the smartphones could continue their and our virtual lives. Here we’re coming to the most conversational ethical problems. What are the data volumes that can be collected and analyzed by the AI? What is, in the end, my smartphone? Is it my virtual avatar that can outlive me and interact with my community? What are the limits?

Not only the ethical or philosophical questions are important, but also the #cybersecurity. What are the data-protection methods? Or can the integrated ChatGPT itself improve the devices’ security? Is it possible, that my command could open me the whole iOS coding system? Can it rewrite the apps or personal data? Maybe even the espionage becomes easier without any specific hacking knowledges (e.g. a jealous partner prompts another partner’s unblocked Siri to send some private information about him/her or likewise could happen with unknown people).

As you can see, I’ve mentioned only a small part of features or problems that can arise with the 苹果 and OpenAI cooperation, but there are still plenty of questions to be discussed and solved. What else could be implemented? What are the main problems and dilemmas?