AI Editorial Policy

Dated: September 10, 2025

Artificial intelligence is a term with many meanings. These days, it often specifically refers to generative artificial intelligence (GenAI) systems, which generate novel text, images or other media in response to user input. What distinguishes GenAI systems from other computer programs is that they aren’t explicitly programmed to accomplish these tasks. Instead, they learn through training on large datasets. This approach has proved powerful, but it also makes the inner workings of GenAI systems opaque. 

Well-known examples of GenAI systems include chatbots such as ChatGPT and Claude, as well as image and video generators. AI transcription, captioning and translation services are based on the same technology, as are many image editing tools, to cite just a few examples. While GenAI has the potential to facilitate and improve newsgathering and storytelling through these tools, it also has the potential to adversely impact the quality and accuracy of science journalism, diminish our credibility and harm our unique relationship with our audiences.

Quanta Magazine has developed a set of principles to guide our use of GenAI. In the interest of transparency, we are sharing these with our audience. 

*************************************************************

Artificial intelligence is not just a potential tool in our newsroom. It’s part of our mission to understand AI, cover advances in the field and explain its origins and impact to our readers. We have been covering AI since well before GenAI tools became widely available, so we understand its capabilities, limitations and potential, and follow changes closely.

As with all the areas that we cover, it is the judgment, curiosity and experience of our writers, editors, artists and producers that make it possible for us to fulfill this mission, by telling the human stories essential to progress in science and math. There are advantages to using GenAI in our processes, but our writers, editors, designers and producers remain directly responsible for everything we create and publish.

Any GenAI use in our research, reporting or writing, production, illustration, photography or video will be done:

With full transparency:

• We will label any substantial use of GenAI in the creation of articles, images, audio, video or any other media. We recognize that “substantial use” is not a specific term, but will disclose creative use of AI when we feel it is meaningful.
• We will also explain how we have mitigated risks, such as the introduction of inaccuracy or bias, with human oversight.
• Where relevant, we will include the name of the GenAI tool used.
• We will not use GenAI to make significant alterations to images, or change an image’s original meaning or intent, without disclosing what changes were made, and what tools were used.

With our mission and values in mind:

• We continue to be committed to producing responsible, freely accessible journalism that is editorially independent, and is meticulously researched, reported, edited and fact-checked through our journalists’ careful vetting, perspective and analysis.
• Our use of GenAI will be consistent with our mission to illuminate basic science and math research through public service journalism.

With our readers in mind:

• We believe that when used responsibly and with the close oversight of our staff, GenAI tools can have the potential to help us deliver our stories more efficiently, and to help us report stories more deeply.
• We also believe GenAI tools have the potential to make Quanta Magazine more accessible to more people through features such as language translations, and through uses of GenAI that we have yet to discover.

With human oversight and accountability:

• All of our content — whether or not it involves GenAI — must meet our editorial standards. It must be accurate, fair and in the public interest.
• We recognize that GenAI tools are prone to bias and error. We will treat any GenAI output as unvetted source material that is subject to human editorial oversight and approval.
• Any use of GenAI in the creation, presentation or distribution of content must include active human editorial oversight and approval.
• We ask contributors to disclose any GenAI use to their editors, and to refrain from sharing any sensitive or confidential information with GenAI tools.

With responsibility for developing competence and literacy in GenAI tools:

• We will use GenAI with the understanding that this is a new, imperfect, and rapidly developing technology.
• We will provide guidance and training for those responsible for GenAI-based decisions.

We plan to reassess our policy on a regular basis as the potential and our use of these GenAI tools continue to evolve.

*************************************************************