[AI Hanip News] Outdated AI Returns as a 'New Model'

Google Assistant and Google Home Evolved with Generative AI

Editor's NoteSince the emergence of ChatGPT, the world has been buzzing with artificial intelligence (AI). Technology is changing rapidly, and news is pouring in every day. Even top industry experts are struggling to keep up. It's overwhelming to consume everything, but you can't afford to miss out either. We deliver the core of weekly AI issues in a concise and easy-to-digest manner.

Smarter AI: From Personalized Answers to Daily Routines

AI services that seemed outdated have returned smarter with generative AI. Google has integrated its existing AI assistant, 'Google Assistant,' with the generative AI 'Bard.' Until now, Google Assistant was limited to simple information like checking the weather or time. So much so that there are even related search queries on how to disable this feature. In contrast, Google Assistant enhanced with Bard can engage in complex and natural conversations. Especially, it provides personalized responses by linking with Google services such as Gmail and Google Docs. For example, if you say, "Organize my business trip schedule to Busan," it will inform you of your daily itinerary and travel routes based on flight and accommodation reservation emails and meeting confirmation emails in your Gmail.

Google Home, which can control various devices in the house such as speakers and home appliances, has also evolved. It added a 'Help Me Script' feature where generative AI writes code for you. You can create customized daily routines with home devices, like "When the fire alarm sounds, blink all the lights in the house and notify the emergency situation through the speakers," or "When a package arrives at the front door, send a notification through the speakers." Even without coding experience, you can input the desired function in text, and it will automatically generate the code.

Help Me Script Demonstration Image [Source=Google Blog]

Generative AI Office Battle

Global software (SW) and service companies such as SAP, Zoom, and LinkedIn have launched AI services one after another. Let's start with SAP, the world's largest enterprise software company based in Germany. SAP unveiled 'Jule,' an enterprise generative AI service. It can generate interview questions, check inventory status, and even help formulate management strategies to overcome poor performance. Zoom, famous for its cloud-based video conferencing service, introduced an AI service called 'Zoom Docs.' Its functions are very similar to AI services from Google or Microsoft (MS), integrating work features such as drafting, editing, and summarizing meetings with the existing video conferencing service. LinkedIn, a social media platform for professionals, supports recruitment tasks with AI. When a recruiter asks for "candidates with over three years of marketing experience in groceries and logistics who can work in Incheon," the AI chatbot provides answers.

Ricky Kapoor, Head of Zoom Asia-Pacific, introducing Zoom Docs [Source=Zoom]

Even Tom Hanks Was Fooled... "Don't Be Deceived by Fakes"

"Don't be fooled by the fake Tom Hanks." This is a warning posted by the famous American actor Tom Hanks on his Instagram. A dental insurance advertisement circulating online features Tom Hanks, but it is a fake created with AI. The Tom Hanks in the ad appears as he did in his younger days, unlike his actual age of 67 this year. American CBS broadcaster Gayle King also experienced a similar incident. An AI-generated image of her was used in a weight loss advertisement without her consent. The global film industry, including Hollywood, is wary of AI-based image synthesis technology. This is because AI can reproduce actors' images and voices instead of hiring actual actors. Hollywood actors' strikes were partly due to these concerns.

Tom Hanks revealed on Instagram that an AI-generated image of himself was used without consent in a dental insurance advertisement.
[Source=Tom Hanks Instagram]

On the other hand, research results have shown that attempts to block AI deepfakes are futile. According to researchers at the University of Maryland, AI-generated images are identified by watermarks, but these can be easily removed. Both visible watermarks and those invisible to the naked eye can be detected and erased with ease. It is also possible to add fake watermarks to images not generated by AI. Researchers from the University of California and Carnegie Mellon University also revealed that simply adjusting image brightness, contrast, or rotating the image can render watermarks useless.

This content was produced with the assistance of AI translation services.