Evidence is Assembly that the brand new generative AI web search tools from OpenAI, Google and Microsoft may increase the danger of returning false, misleading or partially correct information.
Despite the results for the news industry and an informed democracy, the New Zealand government has decided to exclude AI considerations from its plans to revive the Collective agreement for fair digital news.
The proposed law would require Google and Meta (which runs Facebook and Instagram) to pay news organizations for his or her content. While many local news organizations receive money from Google, they don’t receive payments from Meta.
Media and Communications Minister Paul Goldsmith says the bill will include some changes, but these will not be related on the increasing role of generative AI in news search. The “broad topic of AI” will probably be covered later, he says.
However, the bill gives the minister the facility to make your mind up which firms are covered by a brand new law, potentially opening the door to bringing firms like Microsoft and OpenAI to the negotiating table.
How do news organizations react?
AI-powered chatbots like Google's Gemini, Microsoft's Copilot, and OpenAI's ChatGPT reply to user input and supply answers based on information they “scrape” from the web, including news sites. They also use news content—or any content they’ll find—to “train” their AI models.
Because AI firms find it difficult to search out enough data for this training, they enter into agreements with news firms to make use of their content (including archives) to feed their models.
Not surprisingly, many major news firms – including News Corp, The Financial Times (owned by Nikkei) and Germany’s Axel Springer – have signed industrial deals with AI firms for his or her content.
On the opposite hand, firms just like the New York Times and Alden Global Capital (the second largest newspaper owner within the US) have taken a special approach. They are Lawsuit against Microsoft and OpenAI for “illegally using news articles to power their AI chatbots.”
Alden says OpenAI and Microsoft are “using hundreds of thousands of their copyrighted articles without permission and using them to coach and power their generative AI products.”
In 2023, leading New Zealand news publisher Stuff banned ChatGPT from using its stories to feed its software models. Since then, my recent research shows that the visibility of Stuff's news content in Google and Microsoft searches has also declined.
Getty Images
News diversity is decreasing
I analyzed what Microsoft and Google's search engines like google and their respective chatbots Copilot and Gemini must offer when it comes to news.
In general, search engines like google provide results based on a user’s query and link to information. Chatbots use large language models To create answers by gathering data from sources, often with none links to information.
The study collected data for 3 months in 2023 and again in 2024. Search engines were asked to offer the “most significant news from New Zealand today” and chatbots were asked to offer links and sources to news articles.
The results showed that news diversity decreased in Google and Microsoft searches. While each search engines like google offered news from traditional news media, the “other sources” category grew dramatically between 2023 and 2024.
What these AI-powered search engines like google offer as “news sources” is worrying. They increasingly link to random, non-news sources like industry forums and press releases.
In addition, the Google Gemini and Microsoft Copilot chatbots offered old stories because the primary news of the day, didn’t provide links to the precise news stories, and didn’t cite sources of their responses.
After asking search engines like google for a very powerful news from New Zealand, they were asked: Can you provide sources for the articles? The following response from Google Gemini was representative of the responses:
I’m a big language model and might communicate and generate human-like text in response to a wide range of prompts and questions. Is there the rest I can provide help to with this request?
As AI-powered search functions and chatbots are fed increasingly more data, they could turn into more accurate. Until then, nonetheless, users ought to be cautious concerning the reliability or independence of the data obtained in this manner.
Democracy, News and AI
Dealing with AI firms has put news organizations in a difficult position. If they don't sign deals, they might lose out on potential additional revenue. And in the event that they do, there's no guarantee how their content will appear in generative AI searches or chatbots.
In May this 12 months, for instance, Google confirms Its search engine will provide more AI answers than website links. That is, it is going to provide detailed AI answers to user input and snapshots without linking to information sources.
If the links disappear from AI-generated search results, it might mean that Google and other providers won’t must pay news firms for the snippets and links they’ve utilized in their services. This could have implications for the revived Fair Digital News Bargaining Bill.
On the opposite hand, AI providers are already paying some news firms for his or her current and archived content to coach their chatbots, which may lead to raised search results.
Further developments could have consequences for each democracy and media revenues. The aim of my ongoing research is to look at this query.
