House of Lords Communications and Digital Select Committee inquiry: The future of news: impartiality, trust, and technology
EUREKA Robotics Centre is one of the 11 research groups in the UK to access cutting-edge robots and R&D. As the director of the EUREKA Robotics Centre (Esyin Chew) and Social Enterprise Innovator (Shuyang Hu), we are committed to respond the call of evidence for Communications and Digital Committee, UK Parliament about “The future of news: impartiality, trust and technology” based on our 7 years of R&D and teaching experience in Social Media Analytics for Business.
Background to embrace the future: integrating robotics into the UK's legal and ethical frameworks
Some psychologists have suggested that human beings, as products of their environment, are influenced in their judgement by everything that happens around them and the things around them(Ingold 2002). News as an important way for the people to gather information quickly based on trust on the publishers, the people value the impartiality and trust in official press and some social media. Mcluhan (1964) once propounded that “media is the extension of men” and “Media is the Message”, not merely the content of the media it carries. Media influences people's lives and judgements at tall times without huge efforts to validate or facts check. We always hear that the pen of a journalist is menace and thus media. Against this background, it has become particularly important to ensure the context validation and subjective versus objectivity of news, as this directly affects public perceptions and decision-making, using machine learning and based on the premise that the self-publishing profession is on the rise AI and bots may be an option to investigate (Whittaker 2019). They can not only offer assistance in optimising the generation and selection of news content, but also increase the transparency and fairness of news reporting, similarly to the role of black hackers who have criminal intention and white hackers whom the banks employed for enhanced cyber security. The use of robotics and AI in news media is growing, especially in data processing, and planning, scheduling, and optimisation. However, other areas of application in journalism are still not fully developed – the governance. Most AI news projects rely on tech companies (Henriksson, 2023). That is why there is also a need to regulate counterpart companies for complete AI automated or human-AI collaboratively generated news.
Grounded on the past 3 years of thoughtful ethical robotics and AI integration in hospital and care home in both Wales and Malaysia Tokku Zone (a control group) funded by the Welsh Government (Chew, Lee et al. 2021, Yang and Chew 2021), ALTY, 2023), we propound the UK government to establish a national-wide Tokku Zone (Weng, et al., 2015; Shimpo, 2018) where human and robots-AI can co-exist under research and design safely: robotics and AI (including journalism and social media) are given specific legal and operational parameters, with the aim of exploring and refining AI and robotics, and related laws, in a controlled environment. The first two Tokku Zone we declare in Wales is at EUREKA Robotics Centre (2021), Cardiff School of Technologies and Dufrryn Manor Nursing and Residential home (2022) in Wales. The initiative aims to balance innovation with safety, ethical and legal issues, providing unique opportunities for comprehensive research and development. These solutions are critical to maintaining ethical standards and public trust while fostering the same as the journalism industry.
Trends over the next 12 months and 5 years: 1. What impacts (positive and negative) do large technology platforms and online news aggregators have on the UK’s news environment, including media plurality? And how might this change?
Forbes launched a new website called Bertie, an AI content publishing platform powered by a content management system built specifically for in-house newsrooms and partners(Feng and Ots 2018). Bertie generates more compelling headlines, accurately matches images to story content, and assesses reading difficulty. The news site dedicated an article in January last year summarizing the eight tasks that transformer AI can perform for journalists: generating summaries of large texts and documents; generating questions and answers; providing quotes; creating headlines; translating articles into different languages; generating email threads and writing emails; and generating social posts (Kunova, 2023).
Using advanced AI technology can improve the impartiality of news by analyzing and reporting facts more accurately. For example, in a breaking public safety incident, the system was able to collect information from different news sources, analyze live pictures and videos using image recognition technology, supplemented by text analysis to verify the details of the incident and ensure that the information provided was accurate (Gray, Chambers et al. 2012).
By employing AI systems, human bias can be minimized, and news content will be objective. For instance, certain words or phrases which are subjective or show bias can be eliminated from reports, thus supporting accuracy and balance in the content. For instance, AI systems can detect and eliminate the politically oriented statements to maintain without partiality and neutrality the reporting of news. Increase diversity:
AI can gather information from a variety of sources, which results in varied stands, thereby improving the completeness and variability of the news reporting. The Washington Post has created a Heliograf AI tool generating news articles as well as verifying data. A certain Heliograf was employed to track and process social media and vet candidate statements during the 2016 presidential election in the states (Graefe, Bohlken et al. 2020). Applications power by Artificial Intelligence which are dedicated to checking facts can be an ally in the fight against misinformation, as a means through which the news organisations can preserve their rights along with restoring the highest possible Journalistic ethical condition.
Intelligent AI currently producing applications and other generative AI at this stage are essentially just generating data continuously, making probability maximisation, instead of generating logical responses. Due to various flaws in the data and models, generative AI suffers from systematic bias, values confrontation, "viewpoint hegemony", stereotyping, and disinformation. And the models themselves have their limitations. Large-scale language models lack commonsense reasoning, which brings with it limitations in their capabilities. The biggest dilemma with such limitations is that generative AI does not understand the meaning of the text it generates.
To remove bias in news reporting, such as subjective emotions of the contributors or one-sided understanding of the full picture of the event, to increase the credibility of the information, it is important to be objective and neutral in the writing of headlines and content. For example, consider the following news headline: 'A girl suffers harassment incident'. This headline retains the authenticity of the news though. A 19-year-old disabled college student was harassed, and a college student was harassed by an unemployed male. The same three headlines, all of which have journalistic veracity, will be responded to differently by readers who understand the events without preconceived bias. With this example, news organisations should learn how to present events correctly while maintaining the objective judgement of their audience.
How well does regulatory oversight work? Are any changes needed, for example: a) the way Ofcom monitors due impartiality and terms of reference? b) how does Ofcom monitor media plurality?
To ensure that the use of AI in journalism is neither prohibited nor unchecked, it should be regulated through the creation of specific legal restrictions and the establishment of appropriate vetting organisations. This means setting clear frameworks and standards for AI designers, developers, and users to ensure the accuracy and impartiality of content generated by AI programmers. For instance, Ofcom can establish Tokku Zones to trusted and vetted journalists, publishers, and media to integrate safe and ethical AI-Robotics, in conjunction with research groups and centres across the UK. Specifically setting rules and standards, Ofcom could extend its regulatory remit to include setting specific rules and standards for the use of AI in news generation and distribution. This includes requiring truthfulness, objectivity, and diversity in news content, while preventing the dissemination of misleading or biased information.
As well as establishing a vetting body a specialized body or process to receive and deal with complaints about AI-generated content could be established, drawing on Ofcom's mechanism for dealing with public complaints about broadcast content. This body should have both a technical and legal background and be able to effectively assess the compliance of AI content.
Through such measures, combined with Ofcom's existing regulatory framework, it is possible to promote the use of AI technology in journalism while ensuring impartiality and diversity of news content and protecting consumers from misleading content.
Are there any actions the Government should take to address concerns around due impartiality, trust, and the influence of technology platforms? a) Does the Media Bill need to be amended?
At this stage the Media Bill does not have anything to do with the involvement of AI and bots in the creation of news, and limited vetting by AI and bots could be considered in, considering the appropriateness of adding AI and bots for limited vetting. Pilot AI generated and processed news in special areas. From there, AI can be processed for systemic bias, values confrontation, and other troubling issues. Many other countries have published national AI strategy and action plans for addressing this concern, including EU AI Act 2023 (2023), Qatar AI Strategy (2023, pp.12), US AI Bill of Rights (2023) and Malaysia Personal Data Protection Act (2010). The nearest to this strategy in the UK we recommend of is Guidance on AI and data protection (ICO, 2022) and EPSRC Robotics Principles (Bryson, 2017).
News R&D, Governance and industrial organisations can use robotics explored by WT Zone to automate repetitive tasks, allowing journalists to focus on more complex investigative reporting. Instead, they can use robotics and AI algorithms to automatically collect data from the internet, social media platforms, public databases, and specialized news sources. These technologies can monitor and identify news events in real time and automatically collect relevant information, reducing the need for human resources (Opdahl, Tessem et al. 2023). For AI to be sustainable in the news media while fulfilling social responsibility, a clear set of ethical guidelines as well as a legal system needs to be established. These ethical guiding principles should cover truth and accuracy, privacy protection, transparency and interpretation, plurality, and fairness.
To ensure that the information is factual and consequently, to build accountability upon the information the news media must stress the principle of truth and accuracy. Human processes must be strictly audited and verified for AI-generated contents to ensure ethical journalism and legal compliance. Fighting fake news also does not only lead to a more trustful society in terms of news media but to minimising the negative effect of false information in society.
As a rule, the news media should adhere to the principle of user privacy protection under which the news media are supposed to collect, store, and use user data. A strict compliance with relevant laws and regulations pertaining to the protection of personal information at the part of news media will serve to slightly reduce the chances of an invasion of personal privacy. On user data, the users must be clearly told what the data is meant for, and their disclosed categorically obtain consent so that privacy can be protected and misuse or misuse of information in any form is avoided. Transparency and explanation principle is also necessary to stipulate that the news media must comply with the use of AI technology. The human media should bring transparency by sharing with the means and reasons of using AI technologies to the public making all possible to introduce the system’s working principle and eradicate the negative influence of non-explanatory algorithms. A transparent way between allows an advancement of public trust to the news media and readers’ knowledge of the sources and reliability of news stories. To eliminate algorithmic bias and discrimination, the media must actively cooperate to guarantee that news operation is performed with plurality and fairness, acknowledging the truth that different groups’ rights and interests are concerned as well as diversity. Based on principles of plurality and fairness, news media achieve to reach the needs of numerous readers and provide the information that is more comprehensive and objective. Thus, only by observing these ethical guiding principles, the news media can maintain the validity of information, protecting privacy, and impartiality & diversity of content in the appropriate use of artificial intelligence technology. This will allow news media to better accomplish their social obligations, inform the public with credible and useful information, encourage innovation and the progression of their industry towards sustainable development. In the age of AI, by virtue of following the ethical principles, the news media is likely to continue its development to produce more beneficial outcomes for the society.
By adding the step integrating a “Robotic Passport” mechanism into the mix can serve as an innovative solution. This system will serve as a digital book of record; tracking every AI from design to deployment and operational data including identity of creators, purposes for which it is created, ethical compliance of its use and so on. A passport such as these would create openness, enabler regulatory bodies also citizens independently check the ethical practice of AI in news making. They would also allow identification of AI systems, thereby enabling to identify responsible parties for misinformation or biased reporting. This can be supplemented with legislative reforms that outline the explicit criteria that filter AI- generated content and that indicate the privacy standards to be highly respected by the AI applications in journalism maintaining core ethical principles and the accountability standards. The Robotic Passport would, therefore, increase trust, transparency, and accountability in the digital news realm.
As media scholar Joshua Melowitz said: Any media intervention changes the whole aura. Th is does not prevent AI and robotics technologies from supporting and replaced the conventional journalism industry for the current immediate problems with AI. Norms in AI, robotics and their applications are of great importance in journalism to be set up appropriately and effectively as both AI and robotics could be the breaking point for the journalism field to regain the public trust (Chew, 2018).
References
ALTY (2023) ALTY Orthopaedic Hospital, Cardiff University To Establish Malaysia’s First Tokku Zone In A Single-Specialty Hospital. ALTY Orthopaedic Hospital, Cardiff University To Establish Malaysia’s First Tokku Zone In A Single-Specialty Hospital | BusinessToday
Joanna J. Bryson (2017) The meaning of the EPSRC principles of robotics, Connection Science, 29:2, 130-136, DOI: 10.1080/09540091.2017.1313817
Chew, E., et al. (2021). Investigating the first robotic nurses: humanoid robot nightingale and partners for COVID-19 preventive design. International Workshop on Medical and Service Robots, Springer.
Duffryn manor (2022) Healthcare Robotics as a Service (HRaaS), SMART Expertise Project, funded by Welsh Government. Nursing Home - Duffryn Ffrwd Manor Nursing and Residential Care Homes
EU AI Act (2023) EU AI Act: first regulation on artificial intelligence | Topics | European Parliament (europa.eu)
Blueprint for an AI bill of rights (2023) The White House. Available at: https://www.whitehouse.gov/ostp/ai-bill-of-rights/
Malaysia. (2010). Personal Data Protection Act 2010 (Act 709).
EUREKA Robotics Centre (2021) EUREKA Robotics Centre, Cardiff School of Technologies, Cardiff Metropolitan University. www.cardiffmet.ac.uk/eureka
Feng, S. and M. J. J. o. I. A. Ots (2018). "Seeing Native Advertising Production via the Business Model Lens: The Case of Forbes’s BrandVoice Unit." 18(2): 148-161.
Graefe, A., et al. (2020). "Automated journalism: A meta-analysis of readers’ perceptions of human-written in comparison to automated news." 8(3): 50-59.
Blueprint for an AI bill of rights (2023) The White House. Available at: https://www.whitehouse.gov/ostp/ai-bill-of-rights/ (Accessed: 12 February 2024).
Gray, J., et al. (2012). The data journalism handbook: How journalists can use data to improve the news, " O'Reilly Media, Inc.".
ICO (2023) Guidance on AI and data protection | ICO, Information Commission Office.
Ingold, T. (2002). Culture and the perception of the environment. Bush base, forest farm, Routledge: 38-56.
Opdahl, A. L., et al. (2023). "Trustworthy journalism through AI." 146: 102182.
Simpo., F. (2018) Chapter 5: The principal Japanese AI and robot strategy toward establishing basic principles in: Research Handbook on the Law of Artificial Intelligence (elgaronline.com) (pp. 114-142)
Weng, YH., Sugahara, Y., Hashimoto, K. et al. Intersection of “Tokku” Special Zone, Robots, and the Law: A Case Study on Legal Impacts to Humanoid Robots. Int J of Soc Robotics 7, 841–857 (2015). https://doi.org/10.1007/s12369-015-0287-x
Whittaker, J. P. (2019). Tech Giants, Artificial Intelligence and the Future of Journalism, Taylor & Francis.
Yang, J. and E. J. I. J. o. S. R. Chew (2021). "The design model for robotic waitress." 13 (7): 1541-1551.
Chew, E. (2018). Retrieved from https://committees.parliament.uk/publications/written-evidence/
February 2024
8