Lustre

View Original

Ten Things We Learned About ChatGPT.

By Karen and Erica

You’ve probably heard about chatbots, and likely have encountered a few. And by now you’ve surely heard about ChatGPT–a chatbot on steroids that seems to have taken technological advances to a new level. We decided to learn more about it, though we have not been able to use it because there seems to be a long line ahead of us. So–this post was written by us–but maybe when we get access we will have ChatGPT write it again so we can compare.

Meanwhile, here are ten things we have learned that have caused us to think.

  1. Its creators describe ChatGPT as rather friendly:

We’ve trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests. 

2. ChatGPT is smart.

The bot, which was released by the small San Francisco company OpenAI two months ago, amazed users by simply explaining complex concepts and generating ideas from scratch. More important to Google, it looked as if it could offer a new way to search for information on the internet.

3. But ChatGPT is not a genius.

"If you ask it a very well structured question, with the intent that it gives you the right answer, you'll probably get the right answer," said Mike Krause, data science director at a different AI company, Beyond Limits. "It'll be well articulated and sound like it came from some professor at Harvard. But if you throw it a curveball, you'll get nonsense."  

4. And ChatGPT does not know everything–at least, not yet.

[I]nputting “explain how the solar system was made” will give a more detailed result with more paragraphs than “how was the solar system made,” even though both inquiries will give fairly detailed results. You also have the option for more specific inputting requests for an essay with a specific number of paragraphs or a Wikipedia page. We got an extremely detailed result with the request “write a four-paragraph essay explaining Mary Shelley’s Frankenstein.”

If there is enough information available, the generator will fulfill the commands with accurate details. Otherwise, there is potential for ChatGPT to begin filling in gaps with incorrect data. OpenAI notes that these instances are rare. The brand also notes that ChatGPT also currently has “limited knowledge of world events after 2021.” 

5. Nevertheless, at least one Davos participant holds ChatGPT in high regard.

I use it as a writing assistant and as a thought partner.” Wow. A thought partner?

6. ChatGPT is expensive to run. 

ChatGPT had over a million users as of early December — an enviable user base by any measure. But it’s a pricey service to run. According to OpenAI co-founder and CEO Sam Altman, ChatGPT’s operating expenses are “eye-watering,” amounting to a few cents per chat in total compute costs. 

7. People probably will accept the cost, though, given the uses they foresee for ChatGPT, like responding to multiple inquiries that come in all at once—like why is the electricity out. Indeed, huge investments are on the horizon.

8. And then there’s a warfare use case. Of course.

Alex Karp, CEO of Palantir Technologies Inc., a software provider helping governments visualise an army's movements or enterprises vet their supply chains, among other tasks, said such AI could have military applications. Karp told Reuters in Davos, "The idea that an autonomous thing could generate results is basically obviously useful for war." The country that advances the fastest in AI capabilities is "going to define the law of the land," Karp said, adding that it was worth asking how tech would play a role in any conflict with China.   

9. ChatGPT might be a real technological breakthrough, unlike any before. Sounds good except it might result in massive job loss among even highly educated workers.

[AI] really is different, technology experts told [the author]—a range of tasks that up until now were impossible to automate are becoming automatable. “Before, progress was linear and predictable. You figured out the steps and the computer followed them. It followed the procedure; it didn’t learn and it didn’t improvise,” the MIT professor David Autor, one of the world’s foremost experts on employment and technological change, told me. ChatGPT and the like do improvise, promising to destabilize a lot of white-collar work, regardless of whether they eliminate jobs or not.   

10. And some people have concerns about ethical issues, some of which we have already experienced with less advanced technology, like the algorithms that create deepfakes.  

[I] have three predictions. The first is that we will see ChatGPT and tools like it used in adversarial ways that are intended to undermine trust in information environments, pushing people away from public discourse to increasingly homogenous communities. Second, I predict that we’ll see a range of fascinating on-the-ground experiments and research emerging around how we as a society adapt to image and text generation tools like ChatGPT and Dall-E, to use these incredible advances in ways that truly benefit society while limiting harms, particularly to the most vulnerable. Finally, I predict — and hope — we will see growing attention at the federal level to build meaningful guardrails around the development and deployment of these and other AI systems — ones that account for their costs to society and put the protection of fundamental rights and freedoms over pure technical innovation.

So it’s complicated. And also fascinating  

Why are we writing about ChatGPT? Because we want to arm ourselves, and you, with cocktail chatter. And also because we have the sense that this really is a step beyond anything that has gone before. If it is, we don’t think we should rely on government to keep us safe. Or on ChatGPT itself, as the Davos participant might suggest. And it’s not a job just for technology-minded younger people. All kinds of people, but especially people with a few years of experience, need to figure out how AI, and ChatGPT, fit into our world. Because AI is surely here to stay.

It’s an exciting time. Let’s stay involved.