
ChatGPT launched last year with information up to September 2021.
What’s up, folks? Rachyl Jones here, your tech reporting fellow. So, check this out – OpenAI’s got this little weasel running around the internet, gathering fresh info to train its massive language model, ChatGPT. But lately, it’s been hitting some roadblocks.
A bunch of news organizations, like the New York Times, CNN, Reuters, and the Chicago Tribune, have decided to put a lock on their content, denying access to ChatGPT. And get this, even the Australian Broadcasting Corporation and Australian Community Media are keeping their websites under lock and key, according to the Guardian.
Now, this little weasel I’m talking about is actually GPTBot, a web-crawling software that OpenAI rolled out not too long ago. OpenAI says on its website that letting GPTBot have access to your site helps make AI models more accurate and improves their overall capabilities and safety. But, with these news outlets shutting the door on GPTBot, it’s gonna be a bit of an issue for the chatbot known as ChatGPT. It’s missing out on the latest and most relevant news that these media organizations provide.
When ChatGPT was launched, the training only went up until September 2021. So, you know, it’s stuck back in the times of Joe Biden becoming president and the U.S. military pulling out of Afghanistan. In order to stay relevant and on top of the AI game, ChatGPT needs to learn about the new things happening. That’s why OpenAI has been working hard to address this, even releasing a temporary Browse with Bing feature and setting GPTBot free to scour the web. But with the “do not enter” signs put up by these major media organizations, it’s gonna be a challenge for the chatbot.
If you take a look at the terms of service for the Times, Reuters, and the Tribune, they all explicitly state that users are not allowed to scrape their data. The Times goes even further, specifically mentioning that its content cannot be used to train AI programs. At the end of the day, these publishers make their money by selling information, whether through subscriptions or ads. So, freely giving away their content to chatbots, which could potentially eliminate the need for people to visit their websites, would definitely hurt their revenue. And hey, these news outlets are already struggling to figure out their business models ever since social media took away those sweet advertising dollars. The knife is already in their backs, and they’re gonna do whatever they can to keep AI from twisting it.
By blocking GPTBot, these media organizations might just be pressuring OpenAI to pay up for access. Last month, OpenAI struck a deal with the Associated Press to license its news stories for AI training purposes. We don’t know how much OpenAI had to shell out, but you can bet other organizations are gonna be interested in making similar arrangements. Google, for example, is also scraping publishers’ sites to train its own large language model, so it could go ahead and strike deals with news publishers who have locked it out.
Alright, let’s shift gears and see what else is happening in the tech world today.
Rachyl Jones out.
Drop me a line here if you wanna share your thoughts or suggestions with Data Sheet.
NEWSWORTHY
Alibaba’s new A.I. – So, check this out, the Chinese e-commerce giant just launched two new artificial intelligence models, according to CNBC. One of ’em can apparently understand images and even answer basic questions and write captions about ’em. And get this, the other one can engage in complex conversations. Both models are open-source.
Long-awaited regulations for crypto. – The U.S. Treasury Department recently proposed some rules that would treat cryptocurrency platforms kinda like stock brokers, as reported by the Wall Street Journal. This means they’ll have to comply with tax laws and make sure those crypto traders aren’t evading taxes on their earnings. Looks like Congress is cracking down on ’em.
China’s chip imports. – Get this, China’s imports for semiconductor materials have skyrocketed, hitting a record high of $5 billion in June and July. That’s a whopping 70% increase from last year, according to the Financial Times. Now, this surge is happening at a time when the U.S. and its allies are restricting chip exports to China.
ON OUR FEED
“It was time for the EU to set our own rules. A safer Internet for everyone.”
– Thierry Breton, the EU Commissioner for Internal Market, talking about how the Digital Services Act is becoming legally enforceable on Friday for the biggest online platforms. This new law aims to regulate sites that connect people with goods, services, and content, providing greater protections for internet users. It’s part of a bigger wave of regulations aimed at re