Chatbots can now talk, but experts warn they may be listening too

The popular artificial intelligence platform ChatGPT will now have the capability to respond to spoken words and images. This development has raised concerns among some experts who believe that it could result in unwarranted invasions of privacy.


According to a report from the New York Times, OpenAI, the company behind ChatGPT, released the new version of the chatbot on Monday, allowing it to interact with users through spoken language for the first time. Peter Deng, OpenAI's vice president of consumer and enterprise product, stated that the goal is to make ChatGPT easier to use and more helpful.


Similar to well-known platforms such as Amazon's Alexa and Apple's Siri, users of ChatGPT will now be able to converse with the application, which will respond using its own spoken voice. The report also mentioned that ChatGPT will be capable of processing and responding to images. For example, users could upload a picture of their open refrigerator, and ChatGPT would generate recipe ideas based on the contents.


Despite OpenAI's recent efforts to introduce new AI tools, some experts have expressed skepticism regarding the practicality of the updated version of ChatGPT. Christopher Alexander, chief analytics officer of Pioneer Development Group, pointed out that while voice detection and conversational capabilities may seem impressive, they do not significantly enhance the current capabilities of ChatGPT. Essentially, users can simply dictate their Natural Language Processing commands instead of typing them.


Alexander also cautioned that there are considerations to be made for users of the updated platform. He noted that it has the potential to be used as a surveillance tool, as it listens to users and collects additional information. Additionally, he highlighted that as users interact with ChatGPT using their voices, the AI system learns to process voices more accurately in terms of pitch and accent. This training data may enable ChatGPT to develop highly realistic voice capabilities for AI personas in the near future. While this presents creative potential, it could also make it increasingly difficult to detect deep fakes that result from advances in voice technology.


Ziven Havens, policy director at the Bull Moose Project, shares similar concerns and believes that the application could contribute to a worrisome trend of collecting extensive data on Americans. Havens states that as AI continues to advance, companies like OpenAI have the ability to gather even more data, including individuals' voices and images supplied to ChatGPT. Havens calls for congressional action to protect Americans from forfeiting their privacy in the name of innovation.


Jon Schweppe, policy director of the American Principles Project, acknowledges the great creative potential of the new development but warns that it could also facilitate the creation of more deep fakes, making it increasingly challenging to distinguish between AI-generated voices and real human voices. Schweppe's main concern lies in data collection. As ChatGPT is trained to process voices more effectively, it is likely to develop highly realistic voice capabilities for AI personas in the near future.


While users may draw comparisons between ChatGPT's capabilities and those of Siri or Alexa, it is important to note that the platforms rely on different technologies. Alexa and Siri are programmed to perform a set number of tasks or provide specific responses, while ChatGPT utilizes a large language model that can generate additional responses by analyzing vast amounts of internet data.


Phil Siegel, founder of the Center for Advances Preparedness and Threat Response Simulation, believes that these new abilities have the potential to open exciting possibilities for consumers. He describes it as having "an Angel on your shoulder." The ability to engage in spoken conversations and receive timely and useful information, such as recipes, reminders, and information retrieval, makes ChatGPT a compelling consumer application that surpasses the capabilities of Siri or Alexa.