Axios AM

ChatGPT has captured the public imagination in a way the tech world hasn't seen since the iPhone's debut in 2007.

Why it matters: Most of us are only now getting a glimpse of just how smart artificial intelligence has become. It's awe-inducing — and terrifying.

• When ChatGPT launched to the public on Nov. 30, it proved to be much more advanced than even many in the tech industry had expected, Axios' Erica Pandey, Dan Primack and Ina Fried report.

🧐 What it is: ChatGPT is a free (for now) site that lets users pose questions and give directions to a bot that can answer with term papers, sonnets, recipes — almost anything. In almost any style you specify.

The possibilities for ChatGPT seem endless. It recently passed all three parts of the U.S. Medical Licensing Examination (although just barely) as part of a research experiment.

• 🍝 You can tell the chatbot you're a gluten-free lover of Italian food ... and it'll spit out a meal plan and grocery list in seconds.

The array of possible uses is both exciting and scary.

• A chatbot might one day handle complex tasks better than humans. And it could lead us to dark places we can't anticipate.

Between the lines: Researchers have been working on generative AI for a long time. But most of that was outside the public eye. Even some experts were taken aback by how advanced ChatGPT turned out to be.

The bottom line: Artificial intelligence has long seemed like science fiction, or at least like something in the distant future. ChatGPT is forcing us to confront the big role AI may play in our daily lives — and much sooner than we imagined.

• Share this story.

Impressive as ChatGPT is, its current version has severe limitations, as its creators acknowledge.

• The AI tool can put together answers to lots of questions — but it doesn't actually "know" anything, writes Scott Rosenberg, Axios' managing editor for tech.
• That means it has no yardstick for assessing accuracy. It stumbles over matters of common sense.

ChatGPT can't distinguish fact from fiction. It confidently asserts obvious inaccuracies, like "it takes 9 women 1 month to make a baby."

• It "hallucinates" — that is, makes stuff up — at a rate that one expert pegs at 15%–20% of the time.
• It doesn't tell us where it gets information.
• It tries not to provide biased, hateful or malicious responses. But users have been able to defeat its guardrails.

Go deeper.

The rise of generative AI is forcing journalists to reckon with ways automation could improve their jobs — or destroy them.

• For years, automation was perceived as a threat mostly to physical labor and operations work, not creative types. But the latest wave of AI has flipped that script, Axios Media Trends author Sara Fischer reports.

Why it matters: Artificial intelligence has helped automate menial news-gathering tasks like aggregating data. But there's a growing concern that an over-dependence on it could weaken journalistic standards.

CNET last week said it'll pause an AI publishing experiment after being called out for inaccuracies in articles written by AI tools.

• Go deeper.

OpenAI on Monday announced a significant expansion of its partnership with Microsoft, via a "multi-year, multi-billion dollar investment."

• Why it matters: Microsoft long was viewed as the sleepy tech giant, unable to match the innovation and flash of its rivals. That time is over, Axios Pro Rata author Dan Primack writes.

Microsoft's business is productivity. OpenAI is at the forefront of the next wave of productivity tools.

• There's also a good chance that Microsoft will integrate ChatGPT into Bing, which so far has failed to rival Google's search dominance.

Between the lines: Remaining a minority investor lets Microsoft benefit from OpenAI's technology — without taking responsibility for any of its ethical headaches.

• For its part, OpenAI gets billions of dollars without having to run its decisions by Microsoft's ethics or financial committees.

Share this story.

In Davos last week, everyone was trying to deduce AI's meaning for their own little world, reports Axios chief tech correspondent Ina Fried, just back from the Alps:

• CEOs: Eager to seize on the next potential productivity boost, many execs are trying to figure out if tools like ChatGPT can help them cut costs, create new products or get a leg up on the competition.
• Educators: New York City schools were quick to block ChatGPT on their networks — which may turn out to be like trying to prevent students from using calculators.
• National leaders: Who leads in AI matters, Microsoft president Brad Smith said at an Axios event in Davos. "This is going to be extraordinarily important for international competitiveness and the national security of countries," Smith said.

Share this story.

Axios' Ina Fried and Ashley Gold tell us what's coming:

1. The full power of tools like ChatGPT won't be fully unleashed until businesses combine AI with their own data.

• Image-generating tools could help companies create ads at a fraction of today's cost. Movie studios could use the technology to develop an endless array of sequels, spinoffs or games. Today's AI can't complete such projects — but it can provide powerful shortcuts.

2. Regulation around the world will accelerate. There's no AI-specific regulation in the U.S., despite rising attention from federal agencies, the White House and Congress.

3. The hype will get crazy. The import of ChatGPT and generative AI is unquestionable, but the tech industry loves to go overboard.

• With the tech economy in retreat, every struggling company and startup will want to hop on a bandwagon.

Axios Login reader Chris Pulsifer asked ChatGPT to write about itself in Axios Smart Brevity™ style, and shared the results with Ina Fried.

• After a couple of tries and 15 minutes, here's what Pulsifer got:

What's new: ChatGPT is a state-of-the-art language model developed by OpenAI that has been trained on a dataset of over 8 million web pages. It is designed to generate human-like text that can be used to power a variety of natural language processing applications.

Why it matters: ChatGPT has the power to revolutionize the way we use natural language processing in our daily lives and workflows.

Its broad understanding of language allows it to generate relevant and informative text on a wide range of topics.

(That's the edited version. ChatGPT repeats itself a lot.)

How are you using the chatbot? How would you like us to cover it?

• Just hit "reply" to this email, or drop me a line at [email protected].