According to experts, there is a concern that chatbots may be listening as well as talking.
Some experts are worried that the addition of speech and picture recognition capabilities to the popular artificial intelligence platform ChatGPT may lead to unwelcome intrusions into privacy.
The New York Times reports that on Monday, OpenAI, the startup behind ChatGPT, published an updated version of the chatbot that gives it the ability to engage in spoken conversation with users for the first time.
Peter Deng, vice president of consumer and enterprise products at OpenAI, told the New York Times, “We’re looking to make ChatGPT easier to use and more helpful.”
Users of ChatGPT, like those on famous platforms like Amazon’s Alexa and Apple’s Siri, will be able to speak to the program and get spoken responses. According to the article, ChatGPT will also be able to reply to photos, such as users uploading a photo of an open refrigerator to get food suggestions.
Even though OpenAI has been quite proactive in releasing new AI tools over the last several weeks, the revised version of ChatGPT has been met with some skepticism from industry professionals.
Christopher Alexander, chief analytics officer of Pioneer Development Group, told Fox News Digital, “While voice detection and conversational capability sound impressive, they do not really improve the current capabilities of ChatGPT.” “You can skip typing altogether and instead just speak your Natural Language Processing commands.”
Alexander also cautioned that there would be factors to take into account for users of the new platform, stating that it might improve its potential to be used as a “surveillance tool.”
I need to know: WHAT IS CHATGPT?
“You now have ChatGPT listening to you and collecting additional information,” Alexander told the group.
He elaborated, saying that “when you are speaking to ChatGPT, it is learning how to better process voices by pitch, accent, etc.” The voice training that people give to the AI may greatly aid ChatGPT’s future potential for very lifelike speech abilities for AI personalities. This has exciting artistic possibilities, but if it leads to future speech technologies, it might make deep fakes much harder to detect.
The Bull Moose Project’s policy director, Ziven Havens, expressed concern that the app might be a sign of a larger trend toward “collecting unprecedented levels of data on Americans.”
“As AI has developed, so too has the ability of companies like OpenAI to collect even more data,” Havens said, citing the collection of “voices of Americans” and “images” provided by them via ChatGPT. To paraphrase a recent article, “Congress must act to protect Americans from signing away their privacy in the name of innovation.”
Even though the new version has “great creative potential,” American Principles Project policy director Jon Schweppe expressed concern that it “could also open the door for more deep fakes and make it exponentially more difficult to differentiate between AI voice technology and real human voices.”
Obtaining reliable data is a primary problem. “It is likely that in the very near future, ChatGPT will develop incredibly realistic voice capabilities for AI personas as it ‘trains’ to better process voices by pitch, accent, etc.,” Schweppe added.
While the new ChatGPT features may remind some of Siri or Alexa, the article stresses that they are built on distinct technologies. In contrast to the limited functionality of virtual assistants like Alexa and Siri, ChatGPT makes use of a huge language model that may produce extra replies by analyzing massive quantities of data from the internet.
According to Phil Siegel, the founder of the Center for Advanced Preparedness and Threat Response Simulation, such capabilities might give customers access to previously unheard-of opportunities.
Siegel said, “I think this is exactly the type of consumer application that will be useful and have high take-up if designed well.” When it happens, I prefer to think of it as having ‘an Angel on your shoulder.’ One of the most intriguing consumer applications is the ability to have a conversation that yields timely and valuable information (such as recipes, spoken reminders, information retrieval, and others), which paves the way for far more helpful assistants than Siri or Alexa could ever be.