The Associated Press announces guidelines regarding AI usage - and so has TechSpot

[AlphaX]

In context: To say artificial intelligence is a "touchy subject" would be an understatement. While some people view AI as a useful tool that makes everyday tasks easier, others see it as a hindrance. Though AI does have its upsides, it is also not without its flaws. This includes, but is not limited to, inaccuracies, false information, and possible negative effects on employment.

Whether we like it or not, AI is undoubtedly here to stay. As such, many outlets are beginning to make their stance on AI known. One of these is the Associated Press, which recently announced its guidelines on AI usage within any AP material.

Arguably one of the fields most affected by AI is journalism. The Associated Press is one of the biggest news agencies in the world, claiming "4 billion people see AP journalism every day." With such a large number of daily readers, the AP puts an emphasis on ensuring that anything it reports is accurate and unbiased.

Despite its benefits, AI is still only as correct and impartial as the original source it learned from. Due to this, the Associated Press recently declared that journalists may not use AI-generated material in any media published for the news agency. The lone exception to this rule involves stories where the entire premise involves being created by AI.

Along with this, The Associated Press is also updating its Stylebook, which is described as "a roadmap of journalistic practices and rules for use of terminology in stories." AP believes it is essential for journalists to become familiar with AI, as it will almost certainly be something people will be dealing with for years to come, whether positively or negatively.

On a similar note, we at TechSpot recently updated our Ethics Statement.

We have decided to include this regarding our views and usage of AI. The entry reads as follows:

All our editorial content, including news reporting, reviews, tech features, and buying guides, is written by humans. While we embrace new technologies, including AI tools, we don't employ them to generate entirely new text. The use of conventional software to assist with grammar, spelling, and punctuation is standard practice in the industry.

To put it simply, everything posted on TechSpot is researched and written solely by humans. We're currently in a time where false information is more common than ever, and it is important that our readers are able to access truthful and unbiased stories. As far as human error goes, that's entirely possible and inevitable, but still much better than posting unverified or information made up by an AI algorithm.

The Associated Press is not the first news agency to take a stand on AI, and it certainly won't be the last. It's unclear what AI has in store for the future; perhaps it will prove to be a safe source for information down the track. However, recent studies have shown that the complete opposite is currently happening, so only time will tell.

Permalink to story.

https://www.techspot.com/news/99819-associated-press-announces-guidelines-regarding-ai-usage.html

 

[i like foxes]

The level of current AI is low . I ve got the feeling they just collect and provide info but cannot think .

 

[Uncle Al]

Judging by the "qualify" level of currently computer produced news stores, AI can only make it worse ....

 

[cristianm]

I ve got the feeling they just collect and provide info but cannot think .

It surely does not think, it is a generative pretrained transformer. It does not reason. But it can be useful and fun. For sure it should not be used to write news , though it can summarize pretty well and when you are lazy you can even ask it to write something in a good form stating some provided facts. Though, I still think we should not use the results without reading them first

 

[wiyosaya]

On a similar note, we at TechSpot recently updated our Ethics Statement, deciding to include a new addition regarding our views and usage of AI. The entry reads as follows:

• All our editorial content, including news reporting, reviews, tech features, and buying guides, is written by humans. While we embrace new technologies, including AI tools, we don't employ them to generate entirely new text. The use of conventional software to assist with grammar, spelling, and punctuation is standard practice in the industry.

To put it simply, everything posted on TechSpot is researched and written solely by humans. We're currently in a time where false information could be more common than ever, and it is important that our readers are able to access nothing but truthful and unbiased stories.

 

[wiyosaya]

The level of current AI is low . I ve got the feeling they just collect and provide info but cannot think .

AI definitely does not think. While there are some applications where AI has proven useful and beneficial, there are applications where its use is, IMO, appropriate. Just because you can is not a good enough reason to do some things, and that is where the boundary lies, IMO. To ask the question "Would the use of AI in this situation consistently produce beneficial results?" If the answer is no, then I think AI use for what ever the situation, would be inappropriate.

As I see it, AI, used in circumstances beyond which it has already proven very useful, is a glorified search engine replete with all of the search engine flaws.

 

[Dr Roboto]

"While we embrace new technologies, including AI tools, we do not use them to generate entirely new text. Employing conventional software to assist with grammar, spelling, and punctuation is a standard practice in the industry."

The core of TS's statement hinges on "to generate entirely new text." This suggests that while AI tools like ChatGPT may be employed for tasks related to grammar, they're not used for content creation from scratch.

I turn to ChatGPT daily for three specific purposes:
- Assisting with software-related tasks
- Summarizing topics
- Refining and polishing my writing

In my opinion, ChatGPT excels at these tasks. However, one should note that its output may only be around 80% accurate or appropriate. I frequently utilize it to transform my jumble of ideas, musings, and information into polished, professional text. While it's an invaluable tool, it's still essential to review and tweak its outputs. It's an impressive grammar assistant, to say the least.

[disclaimer]This comment "refined" by ChatGPT

 

[VectorVortec]

For me, the worst problem with generative AI is plagiarism. No credit is ever given to the author. The EU is taking a look at this issue.