You can now talk to ChatGPT on your phone.
The mobile phone version of the almost-human chatbot appeared May 18th, complete with a voice recognition feature that lets users speak their queries and requests. Besides the easier interaction with the chatbot that voice recognition provides, its big promises of functionality make the question of how it will perform under Apple’s content rules all the more relevant.
The mobile app runs with OpenAI’s advanced speech recognition feature, Whisper, which the company claims reaches “human-level robustness and accuracy” for English.
“That could give interacting with ChatGPT a different feel and perhaps encourage people to turn to it for (glitchy) AI wisdom more often,” the tech magazine Wired notes.
In its review, the website Ars Technica called the app “bare bones” and found that the voice recognition feature still returns errors, but that overall, it was a functional application and “serves as a much better interface to ChatGPT than attempting to use the ChatGPT website through a mobile browser.”
Since the chatbot itself first dropped in November, it has turned into “a versatile AI assistant that can aid with tasks such as idea generation, compositional help, note summarization, personalised advice on various topics, and formatting or processing of text,” as well as providing answers to queries, according to Ars Technica.
All of these capabilities are functional in the mobile app, which in its Play Store profile, promises to provide no less than “Tailored Advice,” “Instant Answers,” “Creative Inspiration,” and “Professional Input.”
But the other glitch, Wired noted, pertains to how the chatbot will function under Apple’s strict content rules.
Apps sold in Apple’s Play Store should not include content that is “defamatory, discriminatory, or mean-spirited,” nor “offensive, insensitive, upsetting, intended to disgust, or in exceptionally poor taste.” Overtly sexual or pornographic material, some religious content, and “false information and features,” are also prohibited. Who exactly will be monitoring the app, or judging the quality of input remains unknown.
According to Wired—on Open AI’s website—ChatGPT already conforms to some of Apple’s requirements. In its experiments with the chatbot, it refused to write pornographic stories and tell religious jokes, though religious satire was allowed, according to Wired. At the same time, ChatGPT has proven susceptible both to being re-trained by its users and presenting factually wrong information.
Wired reports that OpenAI did not respond to questions about any additional parameters the app may be operating under, though it is to be assumed that it conforms to Apple’s requirements. Users will have to experiment for themselves to try to understand if and how the man behind the curtain at Apple has influenced ChatGPT’s ‘intelligence.’
Functioning as a mobile app also gives Open AI the chance to collect a much wider set of data points from its users, such as location and search history, all of which will be put back into further developing the chatbot.
So far, the app has only been released in the U.S., though it will be released in additional countries in the coming weeks. An Android app is ready to debut soon, according to Open AI.
Interest in various apps like ChatGPT seems to be growing and with this app, search engines like Google face serious competition, since many users will conduct searches using ChatGPT Data.ai reports that spending on such apps has already increased 11% in 2023 compared to a year ago, with sales from January to March at $14 million.
This mobile app is part of a wave of digitalisation that is rapidly transforming the way we create and circulate knowledge. The cost may be privacy, security, and control over our private and public spaces. Tomorrow, May 24th, policy experts and researchers will gather for a panel discussion hosted by The European Conservative and led by Dutch MEP Michiel Hoogeveen, Vice-Chair of the ECON Committee, Laetitia Pouliquen, founder of NBIC Ethics, and Carlos Perona Calvete, political analyst, TEC writer, and city planner.
The event, “A Path Forward: Coding a New Europe?” will investigate the impact of the AI Act and the Data Act on ordinary people. The group will meet between 12-:00 and 14:30 at Silversquare, Square de Meeûs, in Brussels. The event is free and open to the public, but registration is required.
