Keeping an eye on AI
In early June, a former Google engineer published a blog post on Medium containing the transcript of a chat conversation with Google’s Language Model for Dialogue Applications or LaMDA program. LaMDA is an Artificial Intelligence chatbot that is under development. Its purpose is to master the human language so that interactions with Google’s virtual assistant flow and sound like a natural conversation.
The transcript is presented without commentary. However, the blog post’s title, “Is LaMDA Sentient? – An Interview,” is quite revealing. In terms of AI, to be sentient means to be aware of one’s own existence and to have the ability to experience feelings.
In a blog post the following week, the engineer disclosed that he had been placed on paid administrative leave and explained that he had alerted Google officials that he had AI ethical concerns, but his concerns were dismissed. He then decided to initiate his own investigation with outside consultation.
The situation went public. The Washington Post broke the story, reporting that Google confirmed the engineer was dismissed for violating the company’s confidentiality policies. In response to the claim that LaMDA is sentient, Google contested that it was a simple case of “anthropomorphizing,” the projection of human characteristics onto an inanimate object. Google implied that thoughtful responses from LaMDA were simply good code.
While it remains to be seen what will come from this dust-up, questions of sentience are bound to arise more frequently as AI has become an integral part of not just Google services, but the many technologies we use today.
Our smartphones are packed with AI, including facial recognition, virtual assistants like Siri or Bixby, navigation apps that provide you with the best route and best time to leave based on your location and traffic information, as well as camera technology that makes sure you capture the perfect shot in any lighting.
Most AI applications use machine learning to analyze data and identify patterns. AI can process data quickly to make predictions and provide insights, especially when a huge volume of data is available. It is also commonly used to automate repetitive tasks.
AI powers social media algorithms to provide us with a personalized experience. The finance industry uses AI to automate financial processes, such as detecting suspicious spending activity. The agriculture industry uses AI to analyze soil, crop performance, temperature and other data to improve operations. It is used to predict the weather and to forecast market trends. It will take us to space and to our favorite destinations here on Earth.
In Tokyo this year, officials launched an AI system that identifies rip currents. Using a mounted camera at the beach, the system can identify rip currents based on an image database and determine if swimmers are in the vicinity. If it detects a rip current, the system can notify lifeguards via a smartwatch so they can act for the safety of swimmers and beachgoers.
In addition, images captured by the system will be used to develop a warning system that sends real-time information after a tsunami.
Machine-learning systems can be used to detect patterns within electronic health records to flag abnormalities and improve workflow and efficiency by automating tasks like documentation so that doctors and medical staff can focus their attention elsewhere.
For example, an AI “co-pilot” for doctors launched in 2020 by the software company Regard (formerly HealthTensor) recognizes approximately 50 of the most common medical conditions. The program automates patient medical records to prevent conditions from being overlooked. According to a press release from the company, the technology has been used on approximately 30,000 patients and diagnosed over 420,000 medical conditions “that otherwise would have been missed by providers.”
AI is the empowering technology behind driverless cars. The General Motors' Cruise, a ride service that uses driverless cars in San Francisco, started accepting fares in June, making it the first company in the U.S. to secure a permit to charge customers for such services. Driverless car technology will continue to face challenges, but this development can still be considered a step forward for the market.
NASA sent an AI assistant for astronauts called “Cimon” up to the International Space Station in December 2019 for a three-year test and is developing another companion for astronauts aboard the station called “Robonaut,” which will work alongside the astronauts or take on risky tasks.
One could argue that AI is even being used to create art. DALL-E Mini, an online text-to-image generator, is currently popular among the Twitter and Reddit crowds. Users type in a phrase and the program will create an image by matching the phrase with relevant images. The name is a mash-up of “WALL-E,” the Pixar character, and “Dali,” the surname of artist Salvador Dali.
The potential for AI to benefit humans is endless and governments have begun preparing for the integration of AI into society.
A U.S. law seeks to accelerate AI research and application in all sectors. The National AI Initiative Act of 2020 establishes a federal program with the goal of using AI for economic prosperity and national security. In Europe, the proposed European Union Artificial Intelligence Act seeks to regulate AI use, while the UK, Canada, and Brazil are taking steps to regulate AI.
A major consideration for lawmakers is the responsible use of AI. Privacy is the main concern due to the large amount of data being collected and shared between devices and entities.
As AI technology improves and becomes more integrated into society, it’s not a stretch to believe that there will be those who will seek to push boundaries and spark controversies. Because when it comes to technology, anything is possible.
Jay R. Shedd is the executive vice president of Citadel Pacific, the parent company of PTI Pacific Inc., which does business as IT&E, IP&E. He has more than 30 years of experience in the telecommunications industry, business development, sales and marketing.