Articles by: "Rich Murphy, CEO, President and Managing Director – Alliance for Audited Media"

-concept art to show a checklist for an ethical AI framework-

As media companies expand the use of AI to drive efficiency, many are seeking best practices to ensure AI is implemented responsibly and with minimal risk. A recent IAB study found that only one-third of publishers, brands and agencies have adopted formal AI governance tools, which underscores the need for greater accountability and transparency when AI is used.

But where should publishers start? An ethical AI framework provides a blueprint for how organizations can manage risk, protect data and demonstrate compliance when integrating AI into their operations.

Here are examples of each of the eight pillars of ethical AI. When implemented together, these best practices create a strong foundation to help media companies strengthen trust with audiences and partners.

1. Ethical AI Policies

Establishing clear policies that define how AI is being implemented provides the foundation for responsible AI governance. These policies should reflect a commitment to transparency, accountability, minimizing bias and protecting user privacy, and should be reviewed regularly to ensure they evolve with industry developments and regulatory changes.

Example: The Associated Press created a comprehensive policy around its use of generative AI, including who is responsible for monitoring the accuracy of AI output.

2. Transparency and Disclosures

Transparency helps establish greater trust in media. In addition to having an AI policy, publishers should make these policies publicly available and include disclosures when AI is used. A consistent communication strategy, supported by public-facing AI policies, strengthens accountability and promotes confidence among readers, advertisers and partners.

Example: Bay City News, a San Francisco Bay area nonprofit news organization, publicly shares how the team uses AI and adds in-depth context about the process behind each project.

3. Rights and Permissions

It’s also important that media companies secure applicable rights and permissions and the appropriate level of consent to use the information powering AI solutions. This might include establishing internal safeguards to prevent the misuse of proprietary or third-party content. This helps protect media companies from legal challenges and reinforces ethical content development practices.

Example: The New York Times provides guidelines for its staff that copyrighted material should not be input into AI tools, which prevents potential misuse of third-party content in AI training.

4. Accountability and Human Oversight

Since hallucinations and biases in AI exist, human oversight is essential. A “human-in-the-loop” approach ensures that AI outputs have been reviewed before being published. Organizations should assign clear roles for managing AI tools and outputs, and make sure qualified individuals are accountable for how these systems are used.

Example: The USA TODAY Network adds disclaimers to AI-generated article summaries that explain how key points were created and discloses that they were reviewed by a journalist prior to publication.

5. Bias and Fairness

AI models train on existing data, which can include biases that are then amplified. Media companies must implement strategies to identify and mitigate AI bias such as monitoring AI outputs for fairness and inclusivity. Developing a bias mitigation strategy strengthens credibility and supports fair representation across content and audiences.

Example: The Reuters Institute shared several examples from media leaders on identifying AI bias including how The Financial Times’ teams use internal checklists to assess whether AI-generated output from its content recommender system is unbiased and fair across demographics.

6. Privacy and Data Protection

Responsible AI also means demonstrating a strong commitment to protecting data privacy. Companies must ensure that AI systems comply with privacy regulations such as GDPR and CCPA and protect user data. Prioritizing data privacy can reduce legal risk and reinforce a company’s commitment to responsible use of technology.

Example: Graham Media Group, a Detroit-based media company, prioritizes reader privacy and security and shares its compliance with data privacy laws on its disclosure page and in its priva c y policy.

7. Training and Education

Training and education are essential to implementing AI responsibly. Ongoing training ensures that staff understand the capabilities and limitations of AI tools and are made aware of any policy updates as the technology evolves. This includes both general training for all staff as well as proficiency training for individuals responsible for developing, deploying and monitoring AI systems.

Example: Radio-Canada launched a comprehensive AI literacy program to help staff better understand AI and how it should be used. The program includes a foundational training session to provide an overview of AI concepts and ethical considerations as well as follow-up workshops focused on practical applications.

8. Governance and Risk Management

Effective AI governance requires a structured approach to identifying and managing risk. Organizations should embed ethical AI principles into existing policies, ensure legal and regulatory requirements are met and assign clear accountability for oversight. Regular reviews and feedback loops help detect and address emerging risks, allowing companies to adapt as technology evolves.

Example: Gannett’s AI council brings together members from various departments to discuss ideas and assess risk when using AI in new ways.

Responsible AI isn’t about slowing progress but creating a foundation that supports innovation, transparency and trust. Media companies that integrate an ethical AI framework into processes will increase transparency, minimize risk and maintain the credibility that defines trusted media.

Concept art showing areas of AI governance

AI | Perspectives | Supporter Spotlight

From policy to practice: Responsible media AI implementation

June 30, 2025 Rich Murphy, CEO, President and Managing Director – Alliance for Audited Media

As artificial intelligence becomes more embedded in editorial and business processes, media companies face increased pressure to ensure AI is implemented responsibly. This requires companies to develop a plan for AI use that covers several areas including bias mitigation, risk management, legal compliance and long-term governance.

In my last article, I shared real-world examples of how media companies are implementing ethical AI best practices for transparency and disclosures, bias and ongoing staff education. Here we go deeper into the steps media companies are taking to reduce risk, protect privacy and maintain editorial oversight while integrating AI tools into their processes. Together, these form the eight pillars of ethical AI.

Ethical guidelines and standards

Establishing clear policies that define how AI is used across editorial, marketing and operational teams is essential to increasing transparency and building trust with audiences. Already, some media leaders have not only created policies around AI usage but share them publicly – which offer some great examples to other organizations grappling with AI governance.

The New York Times outlines its AI policies as part of its ethical journalism handbook, which was developed for its editorial and opinion teams and is available to the public. The guidelines state the importance of human oversight and adhering to established standards for journalism and editing.

The Financial Times also made its AI governance publicly available by sharing its principles in articles that outline specific tools staff are integrating into their workflows. It also discusses its investment in skill development and how it has transformed into a company committed to AI fluency and innovation.

Media companies need to develop formal AI ethics guidelines that help guide staff and increase transparency with the public. However, it’s equally important to regularly evaluate these guidelines as technology evolves.

Rights and permissions

As part of their governance strategy, companies must also take steps to ensure that any content produced through AI does not infringe on intellectual property rights or violate content licensing agreements. This means securing applicable rights and permissions to use the information generated by the AI tools and creating internal processes to ensure that AI outputs do not use third-party content without permission.

The New York Times encourages staff to use AI to create content including quizzes, quote cards and FAQs. However, its guidelines state that copyrighted material should not be input into AI tools, which prevents potential misuse of third-party content in AI training.

The Guardian outlines its commitment to protecting content creators’ rights when selecting third-party AI tools by stating it will only use tools that have addressed permission, transparency and fair reward for content usage.

These practices can reduce risk and reinforce a publisher’s commitment to responsible content development.

Accountability and human oversight

Even sophisticated AI systems can produce biased, inaccurate or misleading output. To safeguard against this, media companies should take a “human-in-the-loop” approach and assign qualified individuals to oversee AI tools at every stage of use.

Bay City News, a San Francisco Bay area nonprofit news organization, maintains audience transparency by publicly sharing how the team uses AI including in-depth context about the process behind each project. When it created its award-winning election results hub using AI, human oversight including fact-checking was a vital part of the project’s success.

While BBC prohibits the use of AI to directly create news content, AI use in other areas such as research must be actively monitored and the outcomes assessed by an editor.

Wired also does not create content directly from AI, but the company states that if AI is used to suggest headlines or social media posts, an editor needs to approve the final choice for accuracy.

Privacy and data protection

As readers grow more concerned about how their personal data is collected and used, publishers must take steps to ensure that AI tools are deployed in ways that maintain legal compliance. AI governance must include the development of transparent data collection policies and adhering to privacy regulations such as GDPR and CCPA.

Graham Media Group, a Detroit-based media company, prioritizes reader privacy and security and shares its compliance with data privacy laws on its disclosure page and in its privacy policy. The company also uses an in-house AI tool to help employees streamline their workflows without relying on free AI tools or unsecured platforms.

BBC states in its responsible AI policy that if staff intend to include personal data in an AI tool, a data protection impact assessment must be completed prior to use.

Risk management and adaptation

Using AI introduces a range of potential risks such as bias and fairness that must be actively managed. Effective AI governance requires continuous monitoring and a proactive approach to identifying and addressing these risks.

BBC created its AI Risk Advisory Group that includes subject matter experts from legal, data protection, commercial and business affairs and editorial departments. The group provides detailed advice on potential risks of using AI in both the newsroom and across the company.

As AI technologies evolve, so must the ethical frameworks that support their use. By integrating ethical AI principles into daily operations, media organizations can protect their brands, maintain audience trust and demonstrate their value to advertisers and partners who seek reliable, trusted media environments.

AI ethics concept to suggest trust in AI (or not)

AI | Perspectives | Supporter Spotlight

Ethical AI in action: strategies to build audience trust

May 27, 2025 Rich Murphy, CEO, President and Managing Director – Alliance for Audited Media

As artificial intelligence becomes more integrated into both newsroom workflows and business operations, media companies are under increasing pressure to be transparent about how these tools are being used. Recent research shows that most audiences expect clear disclosure when AI plays a role in media production. To maintain trust and credibility, publishers must follow ethical best practices for implementing AI and communicate openly about its use across their organizations.

Here are examples of how several media companies are integrating ethical AI best practices into their operations to build trust and accountability with audiences and advertisers.

Transparency and disclosures

As readers become more aware—and skeptical—of AI’s role in content production, they expect transparency about where and how AI is involved. Developing a disclosure strategy helps audiences better understand the context around AI’s role in producing content.

The Associated Press publicly shares its AI standards and policies and discloses its intent to use AI to improve the quality of its work, while adhering to strict standards for accuracy. The policy also highlights the importance of human oversight in all aspects of content production.

USA Today adds disclosures when AI is used to write summaries or “Key Points” at the top of select articles. The disclosures state that the content was AI-generated and reviewed by a journalist before publishing. USA Today also includes a link to the publication’s ethical conduct policy with the disclosure.

Bias, balance, and fairness

AI models train on a variety of online content. If biases exist in these sources, these biases could be included in AI-generated output. Media companies are creating ways to identify and minimize bias and making these processes public.

The Guardian developed a statement to share how it’s minimizing bias including a working group to discuss ways to prevent bias and develop company-wide policies. The statement also explains how it integrates human oversight into its AI processes.

National Public Radio (NPR) added a special section on AI to its ethics handbook to encourage journalists to identify and counter bias inherent in these tools. The section also emphasizes the importance of providing transparency about how AI is used and discussing questionable output with editors.

Training and education

As AI technologies and best practices evolve, regular training ensures teams stay informed and continue to uphold ethical standards.

Radio-Canada launched a comprehensive AI literacy program to provide a foundation for its staff’s implementation and understanding of AI. The program includes a foundational training session to provide an overview of AI concepts and ethical considerations as well as follow-up workshops focused on practical applications. The company also developed a cross-functional group that includes developers, analysts and journalists to meet bi-weekly to discuss AI developments and share ongoing projects. Since its inception, more than 100 staff members have participated in the training program.

The New York Times implemented a training program that outlines how AI should be used and includes editorial guidelines, use cases and sample prompts. The company also developed an internal team to explore how AI can be applied to newsroom processes.

AI ethics and the broader media ecosystem

Marketers and advertisers are also placing greater focus on ethical standards. The Institute for Advertising Ethics recently released its AI Ethics Toolkit, designed to help buyers navigate the responsible use of AI.

Media organizations can’t slow their pace of innovation. They need to experiment with and implement technologies while ensuring that they keep audience trust at the forefront. The growing focus on ethics across the broader ecosystem signals that publishers who adopt transparent, responsible AI practices will not only build audience trust but also strengthen their position with advertisers seeking accountable media partners.

Advertising | Supporter Spotlight

Media companies need a privacy and data management roadmap

August 5, 2024 Rich Murphy, CEO, President and Managing Director – Alliance for Audited Media

If you find it challenging to keep up with changes in privacy standards and regulation, you’re not alone. Twenty states have passed comprehensive consumer privacy laws, and more are expected to follow. Since a national privacy law does not yet exist, media companies must navigate a patchwork of state and international regulations.

Google’s recent announcement that it will no longer deprecate third-party cookies likely won’t diminish the need for publishers, advertisers and ad tech companies to continue to focus on privacy and protecting consumer data. Google says it will give users the option to consent to tracking. So, some leaders believe this could still lead to deprecation.

With the privacy landscape seemingly always in flux, where do media companies begin? Here’s an overview of what publishers need to know to plan for a compliant future.

Background: Where to start understanding privacy issues

As the digital advertising landscape has evolved, consumers have become increasingly concerned with how their data is collected, used and stored. These concerns led to calls for legislation and industry standards to protect their data and guide businesses on best practices.

The California Consumer Privacy Act (CCPA) was the first state law passed to address this. Key provisions included consumers’ right to know how their data is being used, and the choice to delete or opt out of data collection.

CCPA became a blueprint for other states to follow. But without national legislation, the result has been an assortment of laws that differ across jurisdictions. As advertisers and media companies attempt to navigate these regulations, the industry has recognized the need for solutions.

In March, the IAB released its State of Data 2024 report, which shares feedback from brands, agencies and publishers about current data practices and where the industry is headed. In the study, 95% of respondents expected continued legislation and signal loss this year and beyond. Because of these challenges, companies must significantly change their practices.

Most publishers are taking a state-by-state approach. Brand marketers are opting for a one-size-fits-all strategy aimed at the highest common denominator. Angelina Eng, IAB’s vice president of measurement, addressability, and data center, explained that as organizations navigate a world with greater signal loss, they need to think holistically about the data they’re collecting.

“Companies need to ask themselves who they’re sharing data with, how to activate and whether they have consumer consent. Our research found that nearly 70% of consumers are willing to share their personal data to support advertising overall, and nearly three in four consumers understand that sharing their data enables websites/apps to know more about them in order to serve personalized ads,” Eng said. “We need to provide consumers with guidance and education around the value of allowing advertisers to leverage some data points and provide ads relevant to consumers, which would in turn allow us to measure performance.”

Media industry compliance solutions are in the works

Several industry initiatives are being developed to streamline compliance and make it easier for businesses to implement industry standards.

The IAB Multi-State Privacy Agreement (MSPA) is a framework designed to help companies from all corners of the industry comply with various state-level privacy regulations. It ensures that companies can efficiently manage compliance across jurisdictions.

The Global Privacy Platform works with the MSPA to transmit consumer preferences across jurisdictions, ensuring compliance with privacy laws such as GDPR and state regulations.

Another element of the IAB’s privacy solutions portfolio is its Diligence Platform. This data privacy platform includes standardized privacy diligence questions for different segments of the advertising industry to help streamline the evaluation process and improve compliance efficiency.

Media companies need a roadmap to navigate privacy

As the industry experiences changes including signal loss and new privacy regulations, organizations must build strategies that allows them to continue to leverage first-party data while remaining privacy compliant.

There are new tools to help publishers better understand and navigate these challenges. ThinkMedium, a consulting firm founded by ad tech veteran Dennis Buchheim, recently released its Publisher Readiness Playbook. It Outlines the context, questions and steps for publishers to understand their preparedness for ongoing data- and identity-related shifts.

“Part of the challenge with privacy regulations is that they go beyond laws. It’s the platform policies that in many ways are having a huge impact on the industry,” Buchheim said.

Buchheim added that the number of policies publishers must be aware of can be overwhelming.

“The breadth of regulations and policies is tremendous. You really must understand what applies to you and what doesn’t. You can comply with the strictest interpretation, have a more bespoke plan, or take a blanket approach. Making very conscious decisions like these requires having a good understanding of what’s happening in the industry. We believe the Playbook can help provide some of that understanding.”

Media companies also can gain a better understanding of their level of compliance by participating in an industry certification program. These programs measure companies against current industry standards and can reveal gaps in compliance that may lead to process improvements to ensure they meet industry standards. While keeping up with privacy changes can seem daunting, solutions and guidance exist to help media companies navigate these complexities. Devising a plan and seeking help from industry resources can help media companies remain in good standing with advertisers, consumers and the law.

Advertising | Supporter Spotlight

3 Trust signals that help publishers stand out from MFAs

April 1, 2024 Rich Murphy, CEO, President and Managing Director – Alliance for Audited Media

One topic making headlines in the digital media and advertising industry is made-for-advertising websites. The industry has tried to define them. Studies have been developed to determine how much MFAs are costing advertisers (15% of all ad budgets, according to the ANA). Tech providers have tried blocking them from media buys. Though that has caused some pushback from legitimate publishers’ sites, which might resemble MFAs on the surface (but are anything but). Despite all these efforts, a recent report from Adalytics showed that ads are still reaching MFAs.

The industry is starting to realize that a one-size-fits-all approach to weeding out MFAs doesn’t work. Media buyers and tech providers are taking a step back to re-examine their selection of media partners. Instead of automatically blocking publishers that don’t fit specific criteria, they’re looking at media more granularly and adding approved publishers to inclusion lists.

What are advertisers seeking in a publisher partner? Broadly speaking, they’re looking for media brands that have created quality ad environments that are seen by real audiences. But different buyers use different criteria and technologies, which lead to inconsistent and disparate outcomes. So: How can publishers take control of the narrative and demonstrate why they should be rewarded for their quality content?

The answer: media companies should implement trust signals, especially those developed and backed by the industry. Here are three trust signals designed to help publishers stand out to buyers for their commitment to quality, transparency, and accountability, and earn advertisers’ trust and business.

Trust signal #1: Industry certifications

Several industry groups, including the Media Rating Council, IAB Tech Lab, Trustworthy Accountability Group and the Alliance for Audited Media, have developed compliance programs to recognize publishers for meeting industry standards for ad measurement, brand safety, ad fraud and sustainability. Publishers that complete these programs demonstrate their commitment to transparency. It shows that they have processes and systems in place that deliver accurate and reliable data advertisers can trust.

Many of these industry certification programs require independent validation that a publisher has implemented industry standards and best practices to earn certification. Independent review and validation provide additional assurance for media buyers and publishers.

Trust signal #2: Brand safety tools

With advertisers becoming increasingly concerned about brand safety, solutions have been developed to highlight publishers committed to creating safe content and environments.

The Journalism Trust Initiative was developed by Reporters Without Borders in collaboration with the industry. It provides a blueprint for certifying media outlets for their adherence to industry standards to produce ethical content. JTI’s certification program includes third-party verification and differentiates quality media for their commitment to these standards and best practices.

Trust.txt is an industry framework developed by the non-profit JournalList to help tech platforms and advertisers make associations between publishers and the industry organizations they belong to including Digital Content Next and AAM. Participants in the program add a Trust.txt file to their website directory, which acts as a machine-readable map to help advertisers, tech platforms and search engines connect legitimate publishers with respected industry organizations.

The Ad Fontes Media Bias Chart® rates publishers for their political bias and reliability and lists them on an interactive chart. Publishers verified by a third party tend to be rated among the most reliable. They can use their rating to show media buyers their reliability and trustworthiness. Media outlets rated at the top of the media landscape can take their ratings to the next level and achieve certification from Ad Fontes for consistently providing reliable news and information.

Trust signal #3: Industry memberships

Membership in an industry organization is another signal of trust that differentiates quality publishers from other websites. Industry memberships require investment and often adherence to a code of ethics. Publishers that join and participate in industry associations stand out for their commitment to supporting and advancing the industry.

As digital advertising evolves and becomes more competitive, standing out for your good work becomes even more crucial. Demonstrating your commitment to industry standards and best practices by participating in the industry programs mentioned above is essential for communicating your value to advertisers.

As media buyers become more focused on curating their inclusion lists, it is critical to showcase your commitment to quality and transparency. Adopting these trust signals differentiates premium publishers and valued media brands in a competitive market and fosters a healthier ecosystem for advertisers and audiences alike.