With the "Deus Ex Machina? — Testing AI Tools"-series we want to show you different tools, that aim to simplify writing, design and research by using Artificial Intelligence. More on the "Deus Ex Machina?"-series can be found here.
"AI-powered intelligence for journalists" is what the AI tool JECT.AI is supposed to offer. JECT.AI promises to help users broaden their horizons, adopt new perspectives and thus create more diverse and higher-quality content — especially in the field of science journalism.
To this end, JECT.AI provides users with various application options. Its core element is an AI-based search engine that suggests articles, journalists and scientists for the science topics users searched for, thus providing a broad and detailed overview of current research and reporting on this topic. Five filters allow users to tailor the results to their own needs. In addition to general content, users can search specifically for facts and figures, central characters, background information and developments. For example, if you enter the term "planetary boundaries", the search engine compiles the following results, graded according to related articles and subject areas, scientists and journalists, as well as possible combinations of topics. These results can be further tailored, for example, if a special focus is to be placed on facts and figures.
In addition to the search engine, JECT.AI offers its users five other tools to help communicate science better. Under "Science Audiences", for example, the tool bundles exemplary addressees of science communication, which journalists can use as prototypes for their readers. In addition, common metaphors that reappear in one of the three main science fields that JECT.AI focuses on, storytelling roles, explanations of how science works and indicators of good science journalism can be found among these tools. They are all static content and have not been further modified since the release of JECT.AI.
To use JECT.AI, all you need to do is register using your email address and a password. The basic version is free of charge, but the company offers additional services and extensions on request, which are subject to a fee.
The AI behind the application
JECT.AI emerged as a tool from the EU-funded project: QUEST (QUality and Effectiveness in Science and Technology communication) and is now run as a start-up by its developers Prof. Neil Maiden and Dr. Konstantinos Zachos. The aim of QUEST — which phased out in summer 2021 — was to improve the quantity and quality of science journalism, particularly in the three focus areas of climate change, vaccinations and artificial intelligence.
According to the developers, JECT.AI was developed together with journalists and adapted again and again to fit their needs. Every day, the AI reads more than 1000 newly published articles from over 400 sources (papers, journals, newspapers, etc.) in six European languages (as of 2021) and examines them for their scientific content. Diversity plays an important role in the AI algorithm. It is trained to present the broadest possible spectrum of results. If, for example, 20 scientists are output who are researching a specific area of a topic, JECT.AI ensures that at least 10 of them are FLINTA* (women, lesbians, intersex, non-binary, trans and agender persons). In other words, belonging to groups that are underrepresented in mainstream society and academia.
However, the company does not provide any information on which data set the search engine accesses or whether it has been optimised again since its publication in 2021. The AI on which the search algorithm is based and the data used to train it is also not disclosed on the JECT.AI website.
The rhetorical potential of the tool
JECT.AI is aimed at journalists who are supposed to write about science but have no in-depth scientific expertise. The tool is designed to help them create high-quality content and make science communication more accessible. It helps to find information, knowledge and contacts that facilitate the process of creating informative articles on science topics.
Rhetorically, besides the search engine, which is relevant for finding topics and information, the prepared features are particularly interesting. The personas summarised under "Science Audiences", for example, which represent the addressees of science communication, their attitudes and needs, are intended to help journalists write more specifically for these groups of addressees. These personas indicate, for example, which scientific knowledge, preferred media or forms of interaction apply to certain target groups. Journalists can take these personas as a model and adapt their writing to them so that their communication measures are as effective as possible for the intended target group.
The prepared metaphors and storytelling suggestions also help to build targeted linguistic images and narratives that are sustainable and convincingly convey the scientific content.
The search engine can be used to filter out common topoi and narratives on individual topics. This feature can be instrumentalized to make one's own journalistic contributions more interesting and impressive through new perspectives and more diverse points of view. The possibility to get in touch with other journalists and scientists offers users of JECT.AI the chance to collect and subsequently communicate information of higher quality through exchange. This increases the quality of the content and the logical persuasiveness of the contributions. Especially in subject areas that are less accessible to addressees and require a great deal of expertise (as is usually the case in science), the authority of an expert — such as a scientist — can create trust.
However, the rhetorical potential of JECT.AI remains limited outside of the search engine function. The built-in features offer personas, metaphors and narratives that cover the three scientific fields of vaccination, climate change and artificial intelligence at a basal level, but do not go deeper. There is a lack of terms that are less common and a lack of currency. Since the tools release, these features have not been expanded, which means that changes in discourse, such as new terms or narratives, are not reflected in it. Individually adapting and communicating to a specific audience at a specific time is a core element of any successful communication. Due to the lack of agility and topicality of the tool, exactly this cannot be ensured. The algorithm also does not adapt to the user and sometimes delivers inappropriate results. When searching for the keyword "Friendly AI", for example, the 12 most relevant results returned for this keyword ranged from new developments in artificial intelligence research to an article on football.
Usage in science communication
Basically, JECT.AI is well suited for use in science communication. It can facilitate research processes, establish contacts and, in the process, ensure that science communication becomes a little more diverse and, above all, less presupposed. Theoretically, users do not even need a basic scientific understanding of the tool's focus topics (AI, vaccinations, climate change), but can use JECT.AI to learn the basics of the necessary content, as well as to get help with common formulations within a thematic discourse. In the scientific field of "climate", for example, one finds the metaphor of (rain)forests as the "lungs of the earth". JECT.AI explains the context of the metaphor and how it can be used as accurately as possible in an article or post. At the same time, keywords related to a specific topic are also provided, which can be used to collect and display further information.
The networking potential for journalists and scientists that the application holds is also quite extensive.
The lack of flexibility and breadth of the research areas covered (so far only the three focus topics mentioned above are covered) limits the actual potential of the tool. For young journalists and newcomers to science communication, however, JECT.AI can be a good starting point — albeit exclusively for written science communication, since JECT.AI does not offer more than textual content support. The tool completely excludes creative elements and their communicative potential. The same applies to different media and channels of written science communication. There is no possibility to tailor the results and features to whether, for example, writing is done for newspapers, social media or blogs.
The idea behind JECT.AI to simplify the production of science journalism, provide networking opportunities and map diversity in research are useful starting points for science communication work. The tool offers resources and ideas that can help journalists write differentiated texts, even if they do not have a scientific background themselves or are new to a topic. The tool itself lacks polish, flexibility, individuality adaptability and constant further development of the tools within the tool and the algorithm. Nevertheless, it is a good aid for researching and producing science journalism, but still leaves the main work to the writers. It is suitable for anyone who communicates or is interested in science, regardless of age or level of knowledge.