Ethical AI Principles

Check out our new piece in Nature entitled: "We Need a New Ethics for a World of AI Agents" https://lnkd.in/eSwJCrKu AI is undergoing a profound ‘agentic turn’—shifting from passive tools to autonomous actors in our world. This moment demands a new ethical framework. With Geoff Keeling, Arianna Manzini, PhD (Oxon) & James Evans and the team at Google DeepMind/Google, we focus on two core challenges. 1️⃣ The Alignment Problem: When agents can act in the world, the consequences of misaligned goals become tangible and immediate. 2️⃣ Social Agents: Their ability to form deep, long-term relationships with users introduces new risks of emotional harm. To address this, we must expand our conception of value alignment: It's not enough for an AI agent to simply follow commands. It must also align with broader principles: User well-being, long-term flourishing, and societal norms. For social agents, we argue for an ethics of care: They must be designed to respect user autonomy and serve as a complement—not a surrogate—for a flourishing human life. Moving forward requires proactive stewardship of the entire AI agent ecosystem. This means more realistic evaluations, governance that keeps pace with capabilities, and industry collaboration to ensure this future is safe and human-centric 👍

This is a must read for every HealthTech CEO. The UK Government’s AI Playbook outlines ten principles that ensure AI is used lawfully, ethically, and effectively. 1. Know AI’s Capabilities and Limitations AI is not infallible. Understanding what AI can and cannot do, its risks, and how to mitigate inaccuracies is essential for responsible use. 2. Use AI Lawfully and Ethically Legal compliance and ethical considerations are paramount. AI must be deployed responsibly, with proper data protection, fairness, and risk assessments in place. 3. Ensure Security and Resilience AI systems are vulnerable to cyber threats. Safeguards like security testing and validation checks are necessary to mitigate risks such as data poisoning and adversarial attacks. 4. Maintain Meaningful Human Control AI should not operate unchecked. Human oversight must be embedded in critical decision-making processes to prevent harm and ensure accountability. 5. Manage the Full AI Lifecycle AI systems require continuous monitoring to prevent drift, bias, and inaccuracies. A well-defined lifecycle strategy ensures sustainability and effectiveness. 6. Use the Right Tool for the Job AI is not always the answer. Carefully assess whether AI is the best solution or if traditional methods would be more effective and efficient. 7. Promote Openness and Collaboration Engaging with cross-government communities, civil society, and the public fosters transparency and trust in AI deployments. 8. Work with Commercial Experts Collaboration with commercial and procurement teams ensures AI solutions align with regulatory and ethical standards, whether developed in-house or procured externally. 9. Develop AI Skills and Expertise Upskilling teams on AI’s technical and ethical dimensions is crucial. Decision-makers must understand AI’s impact on governance and strategy. 10. Align AI Use with Organisational Policies AI implementation should adhere to existing governance frameworks, with clear assurance and escalation processes in place. AI in healthcare can be revolutionary if it’s done right. My key (well some) takeaways: - Any AI solution aimed at the NHS must comply with UK AI regulations, GDPR, and NHS-specific security policies. - AI models should be explainable to clinicians and patients to build trust. - AI in healthcare must be clinically validated and continuously monitored. - Having internal AI ethics committees and compliance frameworks will be key to NHS adoption. Is your AI truly NHS ready?

🤖 Would You Want to Be Interviewed by AI? 📰 A new Fortune report just explored this—and the results are shocking. 😲 → 🙅♀️ Candidates find AI interviews cold, impersonal, and frustrating. → 😤 Some even quit halfway—feeling unheard, unseen, and unvalued. → 🚩 Many see it as a red flag for poor company culture. “Job seekers are turning down interviews—because they’re run by AI.” That stopped me cold. I’ve seen AI transform hiring: ⚡ Faster screenings 🎯 Better matches ⚖️ Less bias But if candidates would rather stay unemployed than speak to a bot—we need to pause and ask: what are we missing? Meanwhile, hiring managers are overwhelmed. ⚙️ AI helps them handle thousands of applications, especially for retail, service, and entry-level tech roles. ⏱️ It saves time. ✅ Standardizes screening. 📊 Offers quick decisions. But there’s a growing disconnect: 🤖 AI optimizes for efficiency. 💬 Candidates are craving empathy. In IRREPLACEABLE, I argue we must use AI to augment human judgment—not replace it. Here’s how we can keep hiring human-first: ✅ Use AI only for initial screening—never the full interview. 🤝 Guarantee human contact early. Candidates deserve real interaction. 🧠 Train HR to focus on empathy, context, and culture fit. Because no AI can tell: 🤔 Why did someone hesitate? ✨ What truly motivates them. 🚀 How they’ll thrive in your team. Leaders must ask: are we optimizing for convenience—or connection? AI is here to stay in hiring. But how we use it will define the kind of companies we build. 🏢 💬 Would you walk out of an AI-led interview—or embrace it? Let’s discuss 👇 #AI #Hiring #Leadership #HumanFirst #IRREPLACEABLE #FutureOfWork #Automation #HRTech

Can law help build ethical AI systems by design, or does ethics resist formalization? In earlier posts, I argued that ethics is about reasoned judgement under uncertainty, and that regulation can create clarity where organizations otherwise struggle. With today’s post I want to connect law and ethics to technical implementation; specifically, the role that law can play in facilitating ethical data practices by design. Privacy professionals are well familiar with this concept, as epitomized by Art. 25 GDPR which requires organizations to implement data protection by design and default. But as Prof. Christian Djeffal outlines in a recent article, law by design has since become a fixture of EU law: Law by design translates legal and ethical goals into technical and organizational obligations. At the same time, it deliberately leaves discretion as to implementation. ➡️ What law can do well Frameworks like the GDPR and the AI Act show how law can meaningfully support ethical data practices by design: ✅ They shape how organizations structure the lifecycle of data processing, starting with an initial assessment of the necessity and proportionality of processing. ✅ They require organizations to clearly define roles and responsibilities from the beginning, and document any relevant risks. ✅ They encourage organizations to seek diverse perspectives when developing and deploying new technologies, thus reflecting the inherently interdisciplinary nature of sociotechnical design. ➡️ What this means for ethical AI Ethics is no longer a nice-to-have when it is hardcoded into legal requirements. As I argued in my master's thesis, the AI Act, for instance, translates ethical obligations into technical requirements, specifically mandating: ✅ Respect for human autonomy by requiring human oversight of the development and deployment of AI systems. ✅ The prevention of harm through accuracy, robustness, and security. ✅ Fairness and explainability through robust data governance and record-keeping. ➡️ Where law reaches its limits At the same time, law by design does not resolve any dilemmas or trade-offs. Ethical behavior is not a technological fact, but the result of human deliberation. Procedure matters just as much as outcome, and legal requirements alone do not tell organizations how to weigh competing priorities in practice. ➡️ What this means for leaders on ethical AI Law by design is not a shortcut to ethical AI. But it can create the right incentives. Leaders should: ✅ Leverage law by design requirements as a foundation for responsible data processing. ✅ Facilitate ethical deliberation to translate law by design requirements into concrete deliverables. ✅ Open up the room for innovation by, in Djeffal's words, "prompting the development of solutions where none yet exist." Link to Djeffal's article: https://bit.ly/45Sj76P. #ResponsibleAI #AIGovernance #DataEthics #Leadership

AI regulation isn’t settling, it’s reacting. And the reaction? Fragmented, global and and driven by public tension. Europe: The landmark AI Act is already under review. Why? Industry pushback. Now, the EU is signalling it may ease compliance and reduce red tape. United States: The proposed “AI Diffusion Rule” was pulled just before rollout. The focus has shifted from enforcement to diplomacy. China: Governance is tightening. The details remain unclear, but the intent is unmistakable: more control. It might seem like regulation is shaped only by politics, policy, and industry pressure. But now add the ethical and public concern layer. You don’t need expert analysis. Just read the headlines: →The New York Times is suing OpenAI over training data and copyright boundaries. →A GDPR complaint accuses ChatGPT of generating false, defamatory information. →A U.S. federal judge ordered OpenAI to preserve all ChatGPT outputs, marking a legal shift in how AI content is treated. Three regions. Three agendas. But one emerging pattern: → Public tension surfaces first, whether political, economic, or ethical. → Legal systems scramble to respond. → Governance becomes the tool to contain the risk. So what does this mean for leaders building with AI? If your strategy skips ethical alignment, regulation will catch you off guard. Ethics builds trust. And to navigate today’s grey areas and stay ready for shifting governance, you need to build with adaptability, documentation, and decision traceability in mind. Ethics is the why. Governance is the how. And both are becoming non-negotiable. 👇 How are you preparing for this dual front, ethical accountability and regulatory complexity? Sources in comments

Imagine candidates taking assessments or interviews wearing AI-powered smart glasses that project LLM responses onto lenses only they can see, and/or provide audio responses only they can hear. Are you going to ask candidates to remove their eyewear before an interview? New research from Dunlop & Lievens introduces the FAIR framework - Forbid, Advise, Insulate, Reimagine - and it's one of the most practical models I've seen for thinking through how employers should respond to candidate AI use in hiring. Here's the uncomfortable part: the two defensive strategies - Forbid (detection, proctoring, warnings) and Insulate (controlled settings, face-to-face only) - are explicitly described as temporary. Not potentially temporary. Explicitly temporary. And they make the case convincingly. Agentic AI can now literally take assessments on behalf of candidates - interpreting screens, moving cursors, entering text. Digital proctoring was designed for a world where the cheating tool was a second browser tab, not an agent operating the computer/device itself. But the real insight isn't about technology arms races. It's about construct validity. When some candidates use AI and others don't, your assessment scores aren't measuring what you think they're measuring. You're now capturing an uncontrolled mix of the target skill AND the candidate's GenAI literacy - their ability to recognize use cases, interact effectively with AI, and adapt its output. That's a confound, not a feature. The researchers argue the sustainable path forward is "Advise" and "Reimagine" - equalizing AI access across all candidates and designing assessments where human-AI collaboration IS the thing being measured. Meta is already doing this. Their coding interviews now let candidates use GenAI tools in real time - from a pre-approved set. Think about what that means. Instead of trying to catch people using AI, you're deliberately observing HOW they use it. What questions they ask. How they evaluate and adapt the output. Whether they can push beyond what the AI generates on its own. That's not lowering the bar. That's measuring what actually matters for how work gets done now. If the future of work is human-AI collaboration, then the future of hiring has to assess it too. The question isn't whether your candidates are using AI. They are. The question is whether your assessment strategy is designed around that reality - or still pretending you can prevent it. Links in comments. H/T to my colleague Ellen Whiteside for bringing this research to my attention! 🙏🏻

🔍 Everyone’s discussing what AI agents are capable of—but few are addressing the potential pitfalls. IBM’s AI Ethics Board has just released a report that shifts the conversation. Instead of just highlighting what AI agents can achieve, it confronts the critical risks they pose. Unlike traditional AI models that generate content, AI agents act—they make decisions, take actions, and influence outcomes. This autonomy makes them powerful but also increases the risks they bring. ---------------------------- 📄 Key risks outlined in the report: 🚨 Opaque decision-making – AI agents often operate as black boxes, making it difficult to understand their reasoning. 👁️ Reduced human oversight – Their autonomy can limit real-time monitoring and intervention. 🎯 Misaligned goals – AI agents may confidently act in ways that deviate from human intentions or ethical values. ⚠️ Error propagation – Mistakes in one step can create a domino effect, leading to cascading failures. 🔍 Misinformation risks – Agents can generate and act upon incorrect or misleading data. 🔓 Security concerns – Vulnerabilities like prompt injection can be exploited for harmful purposes. ⚖️ Bias amplification – Without safeguards, AI can reinforce existing prejudices on a larger scale. 🧠 Lack of moral reasoning – Agents struggle with complex ethical decisions and context-based judgment. 🌍 Broader societal impact – Issues like job displacement, trust erosion, and misuse in sensitive fields must be addressed. ---------------------------- 🛠️ How do we mitigate these risks? ✔️ Keep humans in the loop – AI should support decision-making, not replace it. ✔️ Prioritize transparency – Systems should be built for observability, not just optimized for results. ✔️ Set clear guardrails – Constraints should go beyond prompt engineering to ensure responsible behavior. ✔️ Govern AI responsibly – Ethical considerations like fairness, accountability, and alignment with human intent must be embedded into the system. As AI agents continue evolving, one thing is clear: their challenges aren’t just technical—they're also ethical and regulatory. Responsible AI isn’t just about what AI can do but also about what it should be allowed to do. ---------------------------- Thoughts? Let’s discuss! 💡 Sarveshwaran Rajagopal

Your AI isn’t hallucinating. It’s just accurately reflecting your messy data. "There is no AI - without IA." Seth Earley Your Information Architecture (IA) becomes your asset. Like Harari said: "𝙄𝙣𝙛𝙤𝙧𝙢𝙖𝙩𝙞𝙤𝙣 𝙞𝙨 𝙩𝙝𝙚 𝙖𝙩𝙩𝙚𝙢𝙥𝙩 𝙩𝙤 𝙧𝙚𝙛𝙡𝙚𝙘𝙩 𝙧𝙚𝙖𝙡𝙞𝙩𝙮, 𝙩𝙝𝙪𝙨 𝙩𝙝𝙚 𝙩𝙧𝙪𝙩𝙝." If you want your AI solution or Tool to add value to your business (which I think you do) - you need to make sure your model understands your business reality. Your data is that reality. Your IA is the foundation. Here are my 5 Pillars of Data Governance for making data your strategic asset: → 𝟭/ 𝗗𝗮𝘁𝗮 𝗖𝗼𝗹𝗹𝗲𝗰𝘁𝗶𝗼𝗻, 𝗔𝗰𝗾𝘂𝗶𝘀𝗶𝘁𝗶𝗼𝗻 & 𝗥𝗲𝘁𝗶𝗿𝗲𝗺𝗲𝗻𝘁 𝘏𝘰𝘸 𝘴𝘩𝘰𝘶𝘭𝘥 𝘥𝘢𝘵𝘢 𝘦𝘯𝘵𝘦𝘳 𝘢𝘯𝘥 𝘦𝘹𝘪𝘵 𝘺𝘰𝘶𝘳 𝘰𝘳𝘨𝘢𝘯𝘪𝘻𝘢𝘵𝘪𝘰𝘯? - Define legal, ethical, and transparent acquisition channels. - Capture consent and regulatory compliance at source. - Set clear rules for retention and clean, timely deletion. → 𝟮/ 𝗗𝗮𝘁𝗮 𝗦𝘁𝗼𝗿𝗮𝗴𝗲, 𝗢𝗿𝗴𝗮𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻 & 𝗗𝗼𝗰𝘂𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 𝘏𝘰𝘸 𝘥𝘰 𝘸𝘦 𝘴𝘵𝘳𝘶𝘤𝘵𝘶𝘳𝘦, 𝘴𝘵𝘢𝘯𝘥𝘢𝘳𝘥𝘪𝘻𝘦, 𝘢𝘯𝘥 𝘶𝘴𝘦 𝘥𝘢𝘵𝘢 𝘦𝘧𝘧𝘦𝘤𝘵𝘪𝘷𝘦𝘭𝘺? - Data strategy that handles volume, velocity, and variety. - Ensure data marts are business-ready, FAIR, and MECE. - Centralize business rules, logic and KPIs as SSoT. → 𝟯/ 𝗗𝗮𝘁𝗮 𝗤𝘂𝗮𝗹𝗶𝘁𝘆, 𝗢𝘄𝗻𝗲𝗿𝘀𝗵𝗶𝗽 & 𝗦𝘁𝗲𝘄𝗮𝗿𝗱𝘀𝗵𝗶𝗽 𝘏𝘰𝘸 𝘥𝘰 𝘸𝘦 𝘦𝘯𝘴𝘶𝘳𝘦 𝘵𝘳𝘶𝘴𝘵 𝘢𝘯𝘥 𝘢𝘤𝘤𝘰𝘶𝘯𝘵𝘢𝘣𝘪𝘭𝘪𝘵𝘺? - Monitor data accuracy, completeness, and consistency. - Assign clear ownership and stewardship roles. - Establish accountability through data KPIs. → 𝟰/ 𝗗𝗮𝘁𝗮 𝗦𝗲𝗰𝘂𝗿𝗶𝘁𝘆, 𝗔𝗰𝗰𝗲𝘀𝘀 & 𝗣𝗿𝗶𝘃𝗮𝗰𝘆 𝘏𝘰𝘸 𝘥𝘰 𝘸𝘦 𝘱𝘳𝘰𝘵𝘦𝘤𝘵 𝘰𝘶𝘳 𝘥𝘢𝘵𝘢 𝘢𝘯𝘥 𝘴𝘩𝘢𝘳𝘦 𝘪𝘵 𝘳𝘦𝘴𝘱𝘰𝘯𝘴𝘪𝘣𝘭𝘺? - Live data access via “right people, right data, right time”. - Apply anonymization and role-based access control. - Stay compliant (GDPR, HIPAA) and conduct audits. → 𝟱/ 𝗗𝗮𝘁𝗮 𝗨𝘀𝗮𝗴𝗲, 𝗘𝘁𝗵𝗶𝗰𝘀 & 𝗖𝗼𝗺𝗽𝗹𝗶𝗮𝗻𝗰𝗲 𝘏𝘰𝘸 𝘥𝘰 𝘸𝘦 𝘢𝘱𝘱𝘭𝘺 𝘥𝘢𝘵𝘢 𝘪𝘯 𝘱𝘳𝘢𝘤𝘵𝘪𝘤𝘦? - Set clear AI ethics rules, and monitor bias and fairness. - Align with internal policies, laws, and social expectations. - Track data lineage and usage logs for transparency. On a scale of 1 to 10, what priority does Data Governance currently have in your company? 1-3: Data What? 4-7: We're trying, but it's messy. 8-10: It's a strategic pillar. Hi I'm Michael 👨💻 AI Strategist | Keynote Speaker | Executive Coach 👉 Follow to Gain Competitive Advantage through AI

#AI in the public sector? And yet it moves! And it’s a prime example of how technological advancement requires the highest social and ethical standards. “Ethical Integration in Public Sector AI”: the new IAB X Center for Responsible AI Technologies study is out. It addresses the ethical design of AI in the public sector, with a focus on #PublicEmploymentServices (PES). While AI is increasingly employed to streamline administrative processes and improve service delivery, its application in employment mediation raises fundamental concerns regarding #fairness, accountability, and democratic legitimacy. The EU AI Act has further underscored the urgency of addressing these challenges by classifying employment-related AI systems as high-risk. We examine how ethical and social considerations can be systematically embedded in the development and implementation of public sector AI. Using the German PES as a case study, we introduce the “Embedded #Ethics and Social Sciences” approach, which integrates ethical reflection and practitioner involvement from the outset. Qualitative insights from interviews with caseworkers highlight the socio-technical challenges of implementation, particularly the need to reconcile efficiency with citizen trust. We propose concrete design elements emerging from the integration of ethical and social considerations into system development: data ethics, bias, fairness, explainable AI. The approach supports compliance with new regulatory requirements but also strengthens human oversight and shared decision-making.

If your team is asking “Can we use this AI tool?” You need governance.   Especially when AI systems can develop discriminatory bias, give incorrect advice, leak customer data, introduce security flaws, and perpetuate outdated assumptions about users.   AI governance programs and assessments are no longer an optional best practice.   They're on the fast track to becoming mandatory as several AI regulations roll out. Most notably for high-risk AI use. I recommend AI assessments beyond high risk use cases to also capture the privacy, security and ethical risks. Here’s how companies can conduct an AI risk assessment: ✔ Start by building an AI data inventory List every AI tool in use, including hidden ones embedded inside vendor software. Capture data inputs, decisions it makes, who has access, and outputs. ✔ Assess the decision impact Identify where wrong AI decisions could cause harm or discriminate, and review AI systems thoroughly to understand if it involves high-risk.   ✔ Examine company data sources Check whether your training data is current, representative, and free from historical bias. Confirm you have disclosures and permissions for use. ✔ Test for bias and fairness Run scenarios through AI systems with different demographic inputs and look for discrepancies in outcomes. ✔ Document everything Maintain detailed records of the assessment process, findings, and changes you make. Regulations like the EU AI Act and the Colorado AI Act have specific requirements for documenting high-risk AI usage.   ✔ Build monitoring checkpoints Set regular reviews and repeat risk assessments when new products or services are introduced or as models, vendors, business needs, or regulations change. AI oversight isn’t coming someday. It’s here.   Companies that start preparing now will be ready when the new regulations come into force. Read our full blog for more tips and to see how to put this into action 👇