SEO and generative Ai: Inside a “parasitical relationship” //
reports are circulated by AI research lab OpenAI. OpenAI uses news stories from media outlets such as CNN and Wall Street Journal to train its ChatGPT bot. This presents an even greater challenge: How can media outlets maintain traffic, revenue, and relevance in the generative AI age?
Journalists have long been afraid of AI-generated news. The U.K. Press Association, for instance, indicated its intention to use AI in 2016 for certain sports and election stories.
Recent examples of this in the U.S. include the NHL roundup, which was compiled using technology from Data Skrive, a sports content automation company.
Axel Springer Media, which has titles such as Politico and Business Insider, recently stated that AI could replace journalists. Springer wrote to employees, “Only those who create original content will survive.”
“Unknown copyright issues”
The topic of copyrights and possible legal trouble has been raised in France, Spain.
“OpenAI will enhance its model with the most up-to-date information without sending any traffic [to] the original source, it’ll] spark a discussion [over] who has the rights for that content,” stated Marcus Tober senior vice president of enterprise solutions at Semrush.
OpenAI already saw copyright lawsuits and Dan Smullen (head of SEO at Betsperts Media and Technology Group) said that we can expect more soon.
“In fact, despite hearing some publishers adopt AI content within the newsroom, the editorial departments I have spoken with are uncomfortable using OpenAI outputs due to unknown copyright issues,” Smullen said.
OpenAI has taken steps in order to address these concerns. He noted that publishers can opt out of having their content used by OpenAI. AI research lab also agrees to give attribution when it scrapes information from news sites.
Smullen said that SEOs working in the media sector worry that this system might not adequately protect them against copyright or intellectual property issues. News organizations need to continue to monitor OpenAI’s use of news data, and ensure that their content has been used responsibly.
An easy solution is to add footnotes linking directly to sources. This is similar to , which ChatGPT does in Bing.
Smullen said, “We expect something comparable with [Google’s conversational Ai service] Bard.”
‘Truth decay’
AI’s push in news could ultimately lead to a new upsurge in media consumption.
Ben Poulton SEO consultant, founder of the SEO agency Intellar says scraped data can “threaten the curated control news organizations have enjoyed for decades.”
The result could be further degrading of journalistic integrity.
Smullen pointed out that publishers could block OpenAI and its equivalents in the future, which would mean that high-authority news sites will not be crawled. This could lead to a bigger problem with fake news and wider dissemination of biased and inaccurate information that is misrepresented as fact.
Smullen argued that publishers should be compensated for their critical roles. Cameron Conaway is an ex-investigative journalist who heads a growth marketing department at tech giant Cisco. He also teaches digital marketing at San Francisco University.
“Could this increase truth decay and society’s distrust in legitimate new sources?” He asked. He asked, “What effect might this have on democracy if all information is sourceless? And who (or whatever) will hold the power?”
‘Disastrous implications’
OpenAI could eventually be able to automate news production completely, which is a concern. Barry Adams is a specialist SEO consultant at SEO firm Polemic Digital. He noted that generative AI systems cannot predict the news and therefore doesn’t anticipate any immediate problems.
He said that AI will not replace journalism in reporting news, investigating stories, and holding people accountable.
AI could also reword local news stories, without citing as it creates its own versions. Conaway stated that this would lead to a siphoning of traffic and revenue from news sites. This is especially harmful for local news sites that rely on display advertising traffic.
Rewording can also change the meaning of the reporting.
He said that the combination of financially weak and scrappy local newsrooms, media distrust and general media avoidance could have devastating consequences.
It’s not all bad news.
“People will consume news, which is a plus for news organizations. Poulton stated that it’s only the medium that changes. ChatGPT is able to summarize five stories from five different sources in just five seconds. Is that not a great product? ChatGPT might be used on news sites to allow users to quickly find the information they need.
“A parasitic relationship”
The parties must first address the question of traffic and revenue.
Adams stated that he is most concerned about the lack of attribution in early versions of Bing ChatGPT, and Google’s Language Model for Dialogue Applications (or LaMDA), or Bing ChatGPT.
He said that this undermines the fundamental contract of search engines and content websites, which exist in a symbiotic relationship. “GenerativeAI transforms this symbiosis to a parasitical relationship. The search engines take everything (i.e. the content required to train [large-language models (LLMs]) on the web) and return nothing in return.
YouTube, which is owned by Google, uses a more symbiotic approach in which content creators are part of the revenue generated by YouTube.
Adams said that there is no reason why search engines and web could not adopt a similar model, other than the fact that Google would be less money-printing and lose some shareholder value.”
Smullen agreed that the solution was to pay publishers for training data.
Smullen stated that Google will be abused until governments question its legality from a copyright perspective. “It is only fair that publishers are compensated for their contribution to the next generation AI.”
Adams also agreed that it is unlikely that Google will reduce its profits by choice.
He said, “They won’t care that they used a combined knowledge of humanity to build these generative AI system and are now discarding the creators without credit,” he continued. They will do it if they can.
‘Remain vigilant’
According to Julian Scott a content strategist at Socialbu and social media management tool Socialbu, some news organizations have responded by enforcing stricter licensing agreements, imposing stronger data collection and usage rules, as well as using copyright protection software.
He said that these measures might not be sufficient to protect the content’s integrity from misuse without proper attribution.
OpenAI’s model should be improved to ensure proper credit. This is what media industry SEOs want, stated Daniel Chabert CEO and founder at PurpleFire, a web and software development agency.
He said that OpenAI would increase transparency about the use of news data, and be more proactive in notifying authors and publishers when their content has been used.
News organizations should invest in better monitoring systems in order to spot errors and biases in OpenAI’s data.
Chabert said that OpenAI’s use news data should be monitored closely by news organizations and they must take all necessary steps to ensure their accuracy and quality.
‘A first-stop destination’
One tried-and-true online marketing tactic is also available, and it’s particularly applicable here.
Adams said websites should think about the future after Google and create strong brands that connect with their customers.
He said that some publishers have been quite successful in this area and have created brands that are nearly immune to search engine whims. The goal is to be a destination that people visit first, without having to use Google or Facebook.
Matt Greenwood is the SEO manager at Reflect Digital. He said websites should “look to provide information or experiences that are more valuable then can be condensed into just a few lines, to give consumers an incentive to still visit our websites and read our original content.”
Search Engine land published the first News SEO & generative AI: Inside an ‘parasitical relation’.