Engineers are increasingly expected to show evidence that the systems they build in the digital world: AI, social media platforms, and infrastructure systems; are safe, fair and socially responsible. A growing ecosystem of ethics standards is helping them do it.
This week, a US jury set a global precedent by ruling that social media algorithms can be defectively designed and inherently addictive bringing the ethics of engineering into sharp focus.
By holding companies liable for negligence in the design of features like infinite scroll and dopamine-driven recommendation engines, the US legal system has signalled that product liability now extends to the psychological impact of software architecture. This isn’t just about content moderation; it’s about the underlying engineering.
This is the first verdict of many, as 20 more so called bellwether cases await a hearing in US courts, ushering in a new era of liability for software engineers.
These tectonic shifts require new ways of thinking and problem-solving. From AI-enabled systems and smart infrastructure to social media and environmental management, design decisions will now be expected to include bedrock ethical considerations.
A set of standards formulated by the Institute of Electrical and Electronics Engineers (IEEE) guides the way. No longer optional but essential.
A global organisation of about 500,000 professionals, the IEEE in 2021 issued the flagship IEEE7000 standard: Model Process for Addressing Ethical Concerns during System Design. Described as a “set of processes by which organisations can include consideration of ethical values throughout the stages of concept exploration and development”, the standard has been adopted around the world.
It includes steps that organisations of any type or size can take to ensure ethical values are considered in the creation of new systems or technology – from the earliest plans through to final development, and it shows how those values can be tracked as they are built into the system’s goals, design and technical requirements.
Engineer Ruth Lewis MIEAust is Vice President (Standards) of the IEEE Society on Social Implications of Technology (SSIT) and Chair of its standards committee. She worked on the development of the 7000-2021 standard, describing it as consensus-driven and fundamental in showing organisations how to embed ethical principles into the engineering design process from concept onwards.
“Voluntary standards really show the best practice of how to think about these things, particularly in the engineering field or in the IT field, and how much you can incorporate this into your practice to save time and effort,” Lewis told create.
“It’s how you can imbue those sorts of ethical parameters or ethical thinking into that, and apply them to the engineering design process.” Apart from increasing global competitiveness, she added, these standards “make the market move properly”.
Aligning approaches
Now adopted by the International Organization for Standardization (ISO), the 7000-2021 standard gives engineers and technologists an implementable process that shows how to align innovation management processes, system design approaches and software engineering methods, so organisations can address ethical concerns or risks during the design process.
The standard emphasises stakeholder involvement and includes processes to identify the system’s stakeholders throughout its life cycle, and ways to choose stakeholder representatives. The standard requires analysis of the ways the system will be controlled and the collation of relevant data on its social, legal and environmental feasibility.
Reaching agreement on the standard’s process was not always straightforward. “There’s a lot of pressure from the developers,” Lewis said. “Many AI developers fear over-regulation will stifle innovation.”
Overall, the series of IEEE 7000 standards can provide a blueprint for engineering professionals navigating the often uncharted waters of AI. These standards, Lewis added, can assist in determining how best to embed ethical principles in technological projects. These principles include fairness, transparency, inclusivity, privacy and sustainability – built on the understanding that ethical design decisions should have a positive effect for individuals, society and the environment.
Appropriate use
The IEEE began its journey towards the 7000 series standards encompassing ethics, AI and engineering in 2015, with an initiative called Ethically Aligned Design. This initiative is a set of guidelines and principles concerning the creation of autonomous and intelligent systems. Human wellbeing and transparency are prioritised.
The IEEE Principles of Ethically Aligned Design were adopted around the world, including by the OECD and UNESCO.
“It was really the first framing, and it remains also Australia’s AI ethics principles,” Lewis said, noting Ethically Aligned Design is also acknowledged on the Department of Industry website.
The 7000 series of standards includes IEEE 7010-2020, which incorporates an engineering process to assess the impact of AI on human wellbeing, to be used before, during and after design and deployment of any AI system; and IEEE 7007-2021 which defines an ontology for defining ethics in AI and robotics. This standard has been used internationally to define fail-safe mechanisms in lethal autonomous weapons.
Dr Katina Michael, Program Director of the MBA (Technology and Digital Strategy) at the University of Sydney Business School, said standards are vital because they help organisations benchmark against best practice.
“The whole point is to be able to demonstrate that you’ve not only adopted AI or any other technology, but that you’re using it appropriately – and that requires evidence. Policies, stakeholder descriptions, design and development life cycles, user testing results, audit logs, meeting records and any other documentation that could shed light on actual practice.”
Standards emerge when professional bodies identify gaps in the existing landscape, Michael said. They then determine the most appropriate way to address those gaps, whether through a global standard, a national standard or a sector-specific standard.
An increasing number of ethically aligned design standards now cover the growing range of AI and AI-enabled systems in order to identify and mitigate risks.
“There are standards still in their infancy that begin at the industry level, and then gradually rise to national adoption. Those national standards, in turn, feed into the development of international standards,” she said, adding that the field of AI is broad and rapidly expanding, and standards have been developed to keep pace.
“As new developments emerge, different groups need to focus on different elements. We recently saw the release of the AI and procurement standard (IEEE 3119-2025), and there are many other standards that specialise in distinct domains.”
Ethical frameworks and how they are used
| Framework | Origin/Scope | How it’s used by engineers |
|---|---|---|
| Engineers Australia Code of Ethics | Engineers Australia’s national framework for professional conduct | Use as baseline values – integrate ethical questions in design reviews and risk registers |
| IEEE 7000 Series – Ethical System Design | Global standards for AI and autonomous systems | Process model mirroring systems engineering: define stakeholder values, trace ethical requirements, verify compliance |
| IEEE 2089 Series – Age-Appropriate Digital Services | 5Rights principles for children. Age Appropriate Risk Register for Digital Services | Process framework to support the design of “age-appropriate” software and digital services for the wellbeing and flourishing of children online |
| ISO 26000 – Social Responsibility | Global standard for organisational ethics and corporate social responsibility | Use for supply chain and ESG decision audits |
| AS/NZS 5377 – Responsible E-Waste Management | Regional environmental standard | Apply ethical life cycle principles to product end-of-life design |
| EA–IEEE MOU (2025) | Collaboration agreement for shared standards | Enables Australian engineers to adopt IEEE ethical frameworks as practice guidelines |
| UNESCO AI Ethics Recommendation (2021) | UN global policy framework | Inform AI governance within infrastructure and public data projects |
| INCOSE Systemic Risk & Ethics Model (2023) | Systems engineering community guideline | Integrate ethical failure modes into risk analysis and systems resilience planning |
Social media concerns
In 2025, Australia’s eSafety Commissioner cited the IEEE 2089-2021 and IEEE 2089.1-2004 standards in connection with the new social media age-verification regulations, encouraging providers to consider the engineering ethics frameworks.
Since 10 December 2025, a number of social media platforms, including Facebook, Instagram, Kick, Reddit, Snapchat, Threads, TikTok, X and YouTube, have been required to take “reasonable steps” to prevent Australian children under the age of 16 from creating or keeping a social media account. The ambiguity of the term “reasonable steps” is where standards become essential, providing a technical definition for what constitutes a robust age-assurance system.
Michael, the working-party chair for the IEEE 2089-2021 standard which embeds the 5Rights Framework for children, said the standard requires providers to follow best practice in identifying risks and mitigating them through interventions.
“Our hope is that, through the 5Rights principles embedded in the standard, there’s harm minimisation towards eradication,” she said. “It’s placing the onus on the developers who are creating software and systems and services.”
The 5Rights for children are the “right to remove”, so children can edit or delete content they have created; “the right to know”, so children understand who holds or profits from their data and how that data is used; “the right to safety and support”, protecting children from illegal practices; the “right to informed and conscious use”, so children can engage creatively online and have the capacity and support to disconnect; and the “right to digital literacy”, so children are taught how to use, create and critique digital technologies.
The age-appropriate digital risk register in the standard identifies risks early in the development and design phase, so risks can be addressed before implementation, Michael said. These risks include digital services that take the biometrics of a child and study their emotions, in-game purchasing without parental consent, location tracking, and even data privacy breaches.
“Auto-scrolling as a default feature locks young people into additional screen time, behaviourally reinforcing that there is no end-point to a digital service.”
Major international platforms, she added, have already been hit with penalties of hundreds of millions of dollars for breaches in regulation, particularly in the European Union, for breaching children’s privacy.
An eSafety Commissioner spokesperson said the Commissioner encourages the relevant social media platforms to consider current and emerging international standards in terms of deploying age assurance technology, including ISO/IEC FDIS 27566-1 and IEEE 2089.1-2004. Yet alignment or accreditation against a particular standard, the spokesperson said, is not mandatory to comply with the new Social Media Minimum Age legislation.
“Our guidance encourages providers of age-restricted social media to consider relevant international standards and accreditation schemes on a range of matters to inform the taking of reasonable steps,” the spokesperson said. “Alignment with international standards and independent accreditation can help demonstrate that a provider has applied eSafety’s guiding principles in implementing age assurance technology.”
Proof of thought
While alignment is not yet a legal mandate in Australia, the recent US court rulings serve as a cautionary tale for those who treat these standards as optional. By categorising features like infinite scroll as defective designs rather than just policy failures, these rulings have created a direct link between engineering decisions and massive financial liability. For Australian engineers, these standards are no longer just a manual for best practice but the primary defense against the claim that a system was designed without a duty of care.
Ultimately, the shift toward proof of thought means that the engineer’s notebook and the audit log are becoming as important as the code itself. As the legal situation continues to evolve, having a verifiable, standards-based trail of ethical consideration may soon be the only way to prove a system is truly fit for purpose.
Join Ruth Lewis in this on-demand webinar: Navigating the ethics of AI: building skills for responsible engineering.
This is an updated version of the story originally published in the February 2026 edition of create with the headline “Proof of thought”.
