ChatGPT turns 1: Looking back on AI’s breakout year

Before Nov. 30, 2022, interactive AI was, to the average person, pretty primitive.

Alexa and Siri could answer your questions and turn on the lights, but there was no mystery behind the technology: It found answers using search engines and any more complex tasks were programmed. AI art was a novelty. AI completed our sentences and corrected our spelling mistakes. But it hardly seemed like what we were told, through experts’ predictions and media like the 2013 science fiction movie “Her,” AI would be like in the future.

Then came the public release of Open AI’s ChatGPT that November day. Suddenly, everyone could interact with a text bot that could hold a conversation, write a poem or summon a recipe based on the four items in your pantry. It could be trained to be an expert on a certain topic or take on a specific personality.

By February, TIME Magazine declared that human/bot romances were “flourishing.”

“Her” was set in the not-so-distant future. It started coming true almost exactly a decade later.

Looking back

Although it’s the first anniversary of the release of ChatGPT, it’s been a massive year for AI in general. The AI art generator Midjourney, which launched its open beta in July 2022, blew up with its V4 in November 2022 before upgrading to its even more advanced V5 in March 2023. Stable Diffusion, also launched in 2022, upgraded to become SDXL in April 2023, its algorithm advancing staggeringly fast. Now there’s text-to-video.

It didn’t take long for this new world of AI to trigger fear, from concerns that students would use ChatGPT to do their classwork to distress that AI would cause a Terminator-style apocalypse.

In a time of rampant disinformation, there is reason to be concerned. AI may not be able to launch a nuclear war independently, but it can be used as a tool of war.

“Strangely, and for the first time in my history, and I’ve done this for more than 15 years … this is the first time where the senior leaders, who are engineers, have basically said, ‘We want regulation,’” former Google CEO Eric Schmidt said during the Axios AI+ Summit in DC on Nov. 28. “We want regulation, but we want it in the following ways — which, you know, never works in Washington. Nevertheless, we’re doing our best. There is a complete agreement that there are systems and scenarios which are dangerous.”

The regulation question

On Oct. 30, President Joe Biden signed an executive order requiring companies building advanced AI systems to perform safety tests against hackers and share the results with the government before launching products. These guidelines were based on a 1950 law called the Defense Production Act.

According to some experts, applying the laws we already have is as important as creating new regulations. One such commentator is Rohit Chopra, director of the Consumer Financial Protection Bureau who also spoke at the AI+ Summit.

“In many cases, the public has long-standing laws on the books that actually need to be enforced,” Chopra said. “And none of those laws have a fancy technology exception to them.”

Still, AI is the Wild West when it comes to copyright and intellectual property. Artists fear not only that AI bots will take their business, but that generative AI platforms have been trained on their work (and they’re not wrong). Questions like “Who owns a piece of AI art?” have no answer, except that AI art created without human intervention can’t be copyrighted. But is engineering a prompt human engineering? What if a prompt is the literal name of a living artist? What rights do actors have to their own image?

At the same time, some artists see AI as another tool for their toolbox. Iris Creative founder Beth Brodovsky gave a talk called “AI: The Ultimate Sidekick” at the MILLSUMMIT in Wilmington, Delaware last summer. There, she made the case for not fearing AI as an artist, noting that she herself has been a professional illustrator since 1988.

“The world changes, and so I truly believe we cannot live in fear,” she said. “We can’t live in fear that something is going to take our job away. However, if I said I am going to run my career standing at a light table with a Rapidograph pen and drawing black and white drawings, and that’s what I want to do until I die, I would be very hungry right now.”

How we saw AI

Here are a few of AI’s big moments since Nov. 30, 2022, according to Technical.ly’s reporting:

• We have our first experiences with ChatGPT
• Prompt engineering enters the lexicon
• AI chatbots integrate with search engines
• Educators grapple with ChatGPT
• AI enters the workplace
• ChatGPT is asked to help solve social challenges
• ChatGPT drives the angler phishing scam
• AI integrates with Google Docs
• Employers start using AI in the hiring process

Looking ahead

What does 2024 and beyond hold for AI? Expect it to become more prevalent in people’s everyday lives, not less.

AI may be evolving quickly, but it still relies entirely on humans to work. AI sentience is still science fiction, though computers making their own decisions without human input is a future possibility — and some say that’s coming sooner than later.

“Two years ago, [computers setting their own function] was 20 years away. … It had always been 20 years from two years ago,” said Schmidt. “It now looks like it’s 10 or less, from today. And there are people who believe it’s two, three, four years.”

Maybe it’s another Y2K scenario that can be prevented by technologists. Maybe computers will never become independent.

No matter what, it’s going to be a ride.