Artificial intelligence is having a moment.
San Francisco-based research firm OpenAI has made the biggest waves in allowing the public to play with some of its AI tools. ChatGPT is a text-based tool that can carry a realistic conversation as well as write articles, poems and song lyrics. Want instructions on how to make a peanut butter and jelly sandwich in Shakespearean iambic pentameter? No problem.
DALL-E, meanwhile, also accepts natural language, but it takes those descriptions to generate images and art. You can ask it to create a picture of a marble statue of Homer Simpson in the style of Michelangelo, and get something that looks as real as a museum photograph.
There are other tools developing the same capabilities, including Midjourney — an already profitable enterprise that has been tapped to create cover images for The Economist and The Atlantic — and the open-source Stable Diffusion.
Unsurprisingly, there are serious concerns about the ethics and unintended dangers of these tools. While OpenAI insists that it is committed to ensuring that artificial intelligence is safe and “benefits all of humanity,” it doesn’t take much imagination to come up with ways it can be abused.
Critics point to “deepfakes,” or realistic photo or video manipulations that depict celebrities and public figures saying and doing things they would never do — a powerful weapon in this era of widespread coordinated disinformation.
Artists, meanwhile, are crying foul, as the massive computer models that allow AI to write or draw were trained on real writing and real art harvested off the internet. While emulating Shakespeare or Michelangelo might seem harmless, researchers already have found proof that these tools are essentially faking creativity by copying the styles of living humans who already struggle to make a living off their art.
We are again facing a technology that’s developing many times faster than we can truly understand it or control it. For some it’s exhilarating. For others, terrifying.
Lest you think OpenAI has opened a Pandora’s box, however, it’s important to acknowledge that many of us already have let AI into our everyday lives.
Do you ask Apple’s Siri to give you directions, or tell Amazon’s Alexa to start a timer? Do you dictate messages on your phone and marvel at how well — or how poorly — your words are translated into text? That’s artificial intelligence at work.
AI is how websites like Amazon have the uncanny ability to show you things you suddenly need to buy, or how social media apps keep you balanced between curiosity and rage so you keep scrolling. AI is also how doctors can diagnose diseases far earlier and faster than ever before, and how financial institutions and regulators can catch fraud almost instantly.
And one of the most interesting applications of AI can be found in cars, which remain one of the most disruptive innovations of the past century — and one of the deadliest.
Tesla has been the most aggressive in taking experimental tech and putting it on public streets, arguing that fully autonomous cars will eliminate the human errors that kill over 1.3 million people a year. It clearly will be a messy transition, however. Until perfected, self-driving cars will periodically fly off the side of the road or slam blindly into the side of a parked semi.
Meanwhile, on tightly controlled racetracks, scientists are pushing AI driving to extremes, forcing already complex algorithms and control systems to operate at breakneck speeds. And this past weekend, a team of students from the University of Hawaii demonstrated that they are at the cutting edge of AI development.
The UH autonomous race car team, AI Racing Tech, took third place at the Indy Autonomous Challenge, staged at the Las Vegas Motor Speedway. Collaborators included the UH Manoa College of Engineering, UH Maui College, UC Berkeley, UC San Diego and Carnegie Mellon University.
The competition was tough, with only six of nine invited teams getting the opportunity to race. AI Racing Tech made several passes at speeds of more than 125 mph, finishing ahead of a team led by MIT. The triumphant finish comes after the team took second place at a race in November.
Artificial intelligence will inevitably make its way into every facet of modern life. Our UH AI Racing Tech team is demonstrating how AI can make life better — and that Hawaii talent has what it takes to make that happen.
Ryan Kawailani Ozawa is the Pacific news editor for Decrypt, a Web3 media company, and a publisher of the Hawaii Bulletin tech newsletter at HawaiiBulletin.com.
