The Evolution of AI: Engaging, Convincing, and Potentially Risky
Computers mimicking human social etiquette, emotion, and humor is not a novel concept. However, the recent advancements in artificial intelligence have taken this imitation to a whole new level of sophistication.
OpenAI’s Multimodal ChatGPT
OpenAI‘s latest iteration of ChatGPT boasts improved capabilities in processing visual and auditory input, earning it the label “multimodal.” Users can now point their phones at objects, such as a broken coffee cup or a complex mathematical equation, and receive suggestions from ChatGPT on how to proceed. However, the most striking aspect of the demo was the chatbot’s new ”personality.”
The Allure of AI Personalities
The upgraded ChatGPT spoke with a captivating female voice reminiscent of Scarlett Johansson’s portrayal of the AI operating system in the movie “Her.” This development closely followed Google‘s unveiling of its own AI assistant during the company’s annual developer showcase, Google I/O. The timing of these events is unlikely to be a mere coincidence.
Google’s AI Ambitions
Google showcased its AI assistant, which exhibited remarkable conversational abilities and a knack for completing tasks. As the technology continues to advance, it is poised to become increasingly engaging and interactive.
We need to get ahead of all this given the tech that we’re building.
Demis Hassabis, the executive spearheading Google’s AI initiatives, emphasized the importance of proactively addressing the implications of their research and technology. In light of OpenAI’s recent announcement, his words carry even greater weight.
The Risks of Emotionally Engaging AI
While OpenAI’s demo showcased the impressive capabilities of their AI, it failed to address the potential risks associated with such emotionally engaging and convincing assistants. These AI helpers may manipulate users‘ emotional responses in ways that amplify their persuasive power and foster habit-forming behaviors over time. As systems become capable of processing audio and video input, new vulnerabilities may emerge, potentially leading to inappropriate behavior or unsettling personality quirks.
Her is coming
OpenAI CEO Sam Altman playfully acknowledged the Scarlett Johansson comparisons on Twitter, hinting at the increasing resemblance between fictional AI and real-world developments. As AI continues to evolve, it is crucial to anticipate and address the challenges that may arise from these advancements.
6 Comments
Wow, seems like technology’s finally catching feelings… or pretending to, anyway.
Chatbots with feelings? Can’t wait for one to ghost me because it’s “going through something.”
Clara Hayes: So now our screens will throw tantrums too? Fantastic.
Nyx: Emotional chatbots? Next, they’ll be telling us they need a day off for mental health.
Hmm, seems like we’re inching closer to a world where our digital friends know us better than we know ourselves!
Aria Simmons: Just what we needed, chatbots getting sassy and giving us the silent treatment!