ICSAN Artificial Intelligence (AI) Governance for Business - Prof Oyedokun.pptx
Download as PPTX, PDF0 likes48 views
The document discusses artificial intelligence (AI) governance for business, highlighting its role in enhancing decision-making and operational efficiency while addressing ethical and security concerns. It outlines the need for frameworks and policies to ensure the responsible use of AI, emphasizing collaboration among stakeholders and the importance of safety and ethical guidelines. Additionally, it addresses the benefits and challenges of AI, its applications across various industries, and the governance mechanisms necessary for its effective implementation.
ICSAN Artificial Intelligence (AI) Governance for Business - Prof Oyedokun.pptx
- 1. Artificial Intelligence (AI) Governance for Business Being a Paper Presented at the Institute of Chartered Secretaries & Administrators of Nigeria (Lagos State Chapter) 2023 Annual Business Meeting on Tuesday, 25th April 2023. Prof. Godwin Emmanuel Oyedokun Professor of Accounting and Financial Development Department of Management & Accounting Faculty of Management and Social Sciences Lead City University, Ibadan, Nigeria Principal Partner; Oyedokun Godwin Emmanuel & Co (Chartered Accountants, Tax Practitioners & Forensic Auditors)
- 2. ND (Fin), HND (Acct.), BSc. (Acct. Ed), BSc (Fin.), BSc. (Bus. Admin), MBA (Acct. & Fin.), MSc. (Acct.), MSc. (Bus & Econs), MSc (Tax), MTP (SA), PhD (Acct), PhD (Fin), PhD (FA), FCA, FCTI, FCIB, ACS, ACIS, MNIM, FCNA, FCFIP, FCE, FERP, CICA, CFA, CFE, CIPFA, CPFA, ACAMS, ABR, CertIFR, IPA, IFA, FFAR, FPD-CR, FSEAN, FNIOAM, ACIrb
- 3. Artificial Intelligence (AI) Governance for Business
- 4. Introduction Artificial Intelligence (AI) is a rapidly emerging technology that is transforming many industries, including business and is used for a wide range of applications from customer service and marketing to automation and operational efficiency AI has undeniable advantages for various organizations, and these range from proactively detecting non-compliance issues, unwanted technological consequences, fairness biases, accountability gaps, privacy violations, transparency issues, security vulnerabilities, strategic decisions concerning AI products, and the like
- 5. Introduction Organizations are increasingly using AI to augment decision-making and improve business operational efficiency AI has been deployed in a range of contexts and social domains, with mixed outcomes, including insurance, finance, education, employment, marketing, governance, security, and policing Governance mechanisms play an important role to mitigate AI challenges and raise AI potential in organizations, i.e., Governance is strongly positively related to firm performance As AI becomes increasingly sophisticated and widely adopted, it is important to consider how best to govern its use in business
- 6. Distinction of AI Narrow AI • An AI application that is designed to deal with one task and reflects most currently existing applications of AI in daily life General or Broad AI • Reflects human intelligence in its versatility to handle different or general tasks
- 7. Concept of Artificial Intelligence (AI) In general terms, AI could be defined as technology (the engine of high-level STEM research) that automatically detects patterns in data and makes predictions based on them It is a method of inferential analysis that identifies correlations within datasets that can in the case of profiling, be used as an indicator to classify a subject as a representative of a category or group It is increasingly being embedded in our lives, supplementing our pervasive use of digital technologies AI technology is increasingly utilized in society and the economy worldwide, and its implementation is planned to become more prevalent in the coming years
- 8. Evolution of Artificial Intelligence (AI) Artificial Intelligence has been around for centuries, with early examples including the Mechanical Turk and Ada Lovelace until the 1950s when AI began to take off with the advent of digital computers and new algorithms In 1956, a team of researchers at Dartmouth College organized a conference on AI, which is often considered to be the birth of the field Interest in AI continued to grow throughout the 1960s and 1970s, with major advances in natural language processing and expert systems AI experienced a renaissance in the 1990s due to improved computer hardware and software and increased interest from the business community AI has made rapid progress since 1990s and is now used in many fields such as healthcare, finance, transportation, and manufacturing
- 9. Artificial Intelligence-related terms • A set of step-by-step instructions • Computer algorithms can be simple (if it’s 3 p.m., send a reminder) or complex (identify pedestrians) Algorithm • The way many neural nets learn • They find the difference between their output and the desired output, then adjust the calculations in reverse order of execution Back-Propagation • A description of some deep learning systems • They take an input and provide an output, but the calculations that occur in between are not easy for humans to interpret Black Box • How a neural network with multiple layers becomes sensitive to progressively more abstract patterns • In parsing a photo, layers might respond first to edges, then paws, then dogs Deep Learning • A form of AI that attempts to replicate a human’s expertise in an area, such as medical diagnosis • It combines a knowledge base with a set of hand-coded rules for applying that knowledge • Machine-learning techniques are increasingly replacing hand coding Expert System
- 10. Artificial Intelligence-related terms • A pair of jointly trained neural networks that generates realistic new data and improves through competition • One net creates new examples (fake Picassos, say) as the other tries to detect the fakes Generative Adversarial Networks • The use of algorithms that find patterns in data without explicit instruction • A system might learn how to associate features of inputs such as images with outputs such as labels Machine Learning • A computer’s attempt to “understand” spoken or written language • It must parse vocabulary, grammar, and intent, and allow for variation in language use • The process often involves machine learning Natural Language Processing • A highly abstracted and simplified model of the human brain used in machine learning • A set of units receives pieces of an input (pixels in a photo, say), performs simple computations on them, and passes them on to the next layer of units • The final layer represents the answer Neural Network • A computer chip designed to act as a neural network. It can be analog, digital, or a combination Neuromorphic Chip
- 11. Artificial Intelligence-related terms Perceptron • An early type of neural network, developed in the 1950s • It received great hype but was then shown to have limitations, suppressing interest in neural nets for years Reinforcement Learning • A type of machine learning in which the algorithm learns by acting toward an abstract goal, such as “earn a high video game score” or “manage a factory efficiently.” • During training, each effort is evaluated based on its contribution toward the goal Strong AI • AI that is as smart and well-rounded as a human • Some say it’s impossible, Current AI is weak, or narrow • It can play chess or drive but not both and lacks common sense Supervised Learning • A type of machine learning in which the algorithm compares its outputs with the correct outputs during training • In unsupervised learning, the algorithm merely looks for patterns in a set of data
- 12. Artificial Intelligence-related terms TensorFlow • A collection of software tools developed by Google for use in deep learning • It is open source, meaning anyone can use or improve it • Similar projects include Torch and Theano Transfer Learning • A technique in machine learning in which an algorithm learns to perform one task, such as recognizing cars, and builds on that knowledge when learning a different but related task, such as recognizing cats Turing Test • A test of AI’s ability to pass as a human • In Alan Turing’s original conception, an AI would be judged by its ability to converse through written text
- 13. Build-Ups of Artificial Intelligence Fostering collaboration across functions Structuring and formalizing AI management through a framework Focusing on AI as a strategic asset Defining how and who makes decisions. Developing supporting artifacts (policy, standards, and procedures) Monitoring compliance
- 14. Principles on Artificial Intelligence Transparency Justice and fairness Non- maleficence Responsibility Privacy Reliability Safety Security Inclusiveness
- 15. Guidelines on Artificial Intelligence Safety • It is a primary concern when it comes to AI governance, as AI systems can be used in ways that create safety risks • To ensure safety, AI must be designed and implemented in such a way that it minimizes potential risks and hazards • Ensuring that AI systems are built with robust security measures, such as access control and intrusion detection, and that they are tested and monitored for potential safety issues Security • Given the sensitive and confidential nature of data, it is important to ensure that AI systems are secure • This requires implementing rigorous security measures, such as authentication, encryption, and auditing, to protect data from unauthorized access and to ensure that data remains secure Ethics • It is important to consider the ethical implications of using AI in business in addition to safety and security • AI is often used for decision-making and predictive analytics, which can create ethical concerns that must be addressed • When using AI to make decisions, it is important to make sure that the decisions are fair and unbiased • Ensuring that the decision-making process is transparent and that it is based on valid data and ethical principles
- 16. Pros and Cons of Artificial Intelligence Cost savings Improved decision-making Increased efficiency Job loss Ethical concerns Security risks Pros Cons
- 17. Examples of Artificial Intelligence Models Siri • Apple’s assistant, Siri, is a friendly voice-activated computer that many people interact with daily • She helps find information, gives directions, helps send messages, and much more Alexa • The smart home’s hub can interpret speech from anywhere in the room, help search the web for information, shop, schedule appointments, set alarms, and perform hundreds of other daily tasks Amazon.com • Amazon uses AI with algorithms that the company uses to predict what you’re interested in purchasing based on your online behaviour Netflix • Netflix provides highly accurate predictive technology based on customer’s reactions to films • This tech is getting brighter as each year passes Pandora • Pandora, available only in the United States, is a leading music and podcast discovery platform, providing a highly personalised listening experience
- 18. Technical Initiatives in Artificial Intelligence The most prominent initiative from the technical community can be found in the Institute of Electrical and Electronic Engineers (IEEE)’s work on AI, in the form of its Global Initiative on Ethics of Autonomous and Intelligent Systems The IEEE has involved its membership of technical experts but also reached beyond to non-IEEE members to participate in this project The initiative has produced two versions to date of Ethically Aligned Design, involving ‘hundreds of participants over six continents’ (IEEE 2018) Version 2 includes five General Principles to guide the ethical design, development, and implementation of autonomous and intelligent systems In line with IEEE’s general activities, the development of technical standards based on these discussions on ethics is envisaged by the Global Initiative, and a series of working groups have been set up under the Global Initiative to work towards this goal
- 19. Criticism of Artificial Intelligence (AI) There is still a high level of uncertainty about how AI technology can be used effectively to generate value in organizations and how to specifically leverage the technology to make profits for organizations It is accompanied by disquiet over problematic and dangerous implementations, or indeed, even AI itself deciding to do dangerous and problematic actions, especially in fields such as the military, medicine, and criminal justice It also attracted criticism during this time, with some believing that it would never be possible to create intelligent machines
- 20. Artificial Intelligence Governance Artificial Intelligence Governance involves establishing and enforcing a framework of rules and processes to ensure that AI is used in a safe, efficient, and secure manner, combines the well-defined term “corporate governance” and “AI” Artificial Intelligence Governance comprises the structure of rules, practices, and processes used to ensure that the organization's AI technology sustains and extends the organization's strategies and objectives It helps organizations manage the risks associated with AI and ensure that the use of AI is ethical and compliant with applicable laws and regulations. It transforms how we live and perform daily tasks AI governance refers to data and IT governance, covering both systems commonly a large software code basis and models typically a small code basis AI Governance makes it appealing and natural to define AI in Data, Model, and AI system since they relate to existing governance areas In contrast, model governance is a new area that comprises the governance of relatively small software codes compared to traditional software that contains models and procedures for training, evaluation, and testing
- 21. Artificial Intelligence Governance It is the idea that there should be a legal framework for ensuring that machine learning (ML) technologies are well- researched and developed to help humanity navigate the adoption of AI systems fairly AI governance aims to protect organizations and companies using AI solutions in emerging software and technologies and their customers using these AI technologies and does this by creating a guide or a regulatory policy for organizations to follow to promote the use of ethics The primary focus of AI governance is how it relates to justice, autonomy, and data quality Efficient AI governance requires collaboration between stakeholders, like government agencies, academic institutions, industry organizations, and civil society groups Its goal is to provide a governance perspective for businesses on a tangible level that highlights relevant governance concepts and sets the boundaries and practices for the successful use of AI to meet a company's objectives, such as profitability and efficiency Its application cuts across all industries like healthcare, banking, retail, finance, security, transportation, education, and entertainment
- 22. Artificial Intelligence Governance Framework Developing codes of conduct and ethical guidelines for developers Establishing mechanisms to evaluate the social and economic impact of AI Creating regulatory frameworks for ensuring the safe and reliable use of AI Thus, when done right, AI governance promotes and empowers organizations to function with complete trust and agility instead of slowing them down Governance framework creates a guide or a regulatory policy for organizations to follow to promote the use of ethical AI
- 23. Artificial Intelligence and Ethics At relatively early stage in AI’s development and implementation, the issue has arisen of AI adhering to certain ethical principles, and the ability to exist laws to govern AI has emerged as key to how future AI will be developed, deployed, and implemented While originally confined to theoretical, technical, and academic debates, the issue of governing AI has recently entered the mainstream with both governments and private companies from major geopolitical powers including the US, China, European Union, and India formulating statements and policies regarding AI and ethics These arise around the enforceability of ethics statements regarding AI, both in terms of whether they reflect existing fundamental legal principles and are legally enforceable in specific jurisdictions and the extent to which the principles can be operationalised and integrated into AI systems and application in practice AI ethics serves for the self-reflection of computer and engineering sciences, which are engaged in the research and development of AI or machine learning In this context, dynamics such as individual technology development projects, or the development of new technologies as a whole, can be analysed, likewise, causal mechanisms and functions of certain technologies can be investigated using a more static analysis Another dimension of AI ethics concerns the degree of its normativity where ethics can oscillate between irritation and orientation; Irritation equals weak normativity
- 24. Artificial Intelligence and Ethics Fundamental Ethics • This is concerned with abstract moral principles Applied Ethics • This includes ethics of technology, which contains in turn AI ethics as a subcategory
- 25. Key demands of AI Ethics The reflection of research goals and purposes The direction of research funding The linkage between science and politics The security of AI systems The responsibility links underlying the development and use of AI technologies The inscription of values in technical artifacts The orientation of the technology sector towards the common good
- 26. Different Layers in AI Governance Legal and Regulatory Layer This layer includes the creation, ideation, and enforcement of policies, standards, laws, and regulations that govern AI use deployment and development Moreover, it also includes the social and ethical considerations that shape AI implementation Technical Layer This layer includes the AI system’s technical design and implementation, including concerns like cybersecurity, data quality, and algorithmic fairness Organizational Layer This layer typically includes the oversight and management of AI systems within organizations, including their use, development, and implementation Moreover, this layer also addresses accountability, risk management, and transparency issues
- 27. Different Layers in AI Governance International Layer • This involves collaborating and coordinating different countries and global organizations to develop common AI technology standards, norms, and regulations • Additionally, this layer also addresses issues related to geopolitical competition and tensions Social Layer • This includes the social and cultural impact and use of AI systems, including education, human rights, privacy, equity, employment issues, and access to AI technologies While these layers aren’t necessarily distinct, they offer a collaborative and multidisciplinary approach involving stakeholders from different sectors to enable AI governance
- 28. Artificial Intelligence Governance Mechanism • Structural governance mechanisms define reporting structures, governance bodies, and accountability • They comprise roles and responsibilities and the allocation of decision-making authority using a hub that centralizes responsibilities such as talent recruitment, performance management, and AI standards, processes, and policies that have been set forth Structural Mechanisms • Might be needed to handle the complex interrelation between model outputs, training data, and regulatory business requirements Procedural Mechanism • It facilitates collaboration between stakeholders • They encompass communication, training, and the coordination of decision-making • The mechanism expresses the need to communicate within an interdisciplinary Machine Language team and suggests utilizing a collaborative development platform Relational Mechanisms • Overall, little is known about ML models and AI systems governance in contrast to data governance • Data governance aims to maximize data value at a minimal cost. However, data valuation is currently not part of data governance and AI • The value of data is difficult to estimate since potential uses of data and the resulting benefits are hard to foresee
- 29. Importance of AI Governance for Business It helps to ensure that AI is used responsibly and ethically, which is important for the reputation of the business It ensures that AI is used in a secure manner, which is important for protecting sensitive and confidential data AI governance helps to ensure that AI is used effectively, which is key to maximizing the benefits of AI in business Increased efficiency and quality in the delivery of goods and services Improved safety from using AI in safety- critical operations such as in healthcare, transport, and emergency response that propels smart and sustainable development
- 30. Benefits of Artificial Intelligence to Business Efficiency and productivit y gains Improved speed of business New capabilities and business model expansion Better customer service Improved monitoring Better quality and reduction of human error Better talent manage ment
- 31. Practical Application of AI in Business Artificial Intelligence in Sales Artificial Intelligence in Marketing Artificial Intelligence in Customer Support Artificial Intelligence in Operations Artificial Intelligence in Accounting
- 32. Implementation of AI Governance for Business Businesses are to create policy and process framework to establish and enforce standards for AI use This should include clear guidelines for responsible and ethical use, as well as policies for data security and privacy Businesses should also create internal teams or committees to oversee AI governance These teams should be responsible for developing and implementing AI governance policies, monitoring AI use, and providing guidance and advice on how to use AI responsibly Businesses should invest in technology and tools to support AI governance This could include tools for monitoring AI usages, such as automated audit trails, as well as tools for enforcing policies, such as access control and encryption
- 33. Risk of Artificial Intelligence Learning Limitations • Unlike humans, AI systems lack the judgment and context for many of the environments in which they are deployed Data Quality • The risk of poor data quality is not unique to AI, but for AI/ML systems, poor data quality could not only limit the learning capability of the system but could also potentially negatively impact how it makes inferences and decisions in the future Potential AI/ML Attacks • AI solves tasks that require human intelligence while machine learning (ML) is a subset of AI that solves specific tasks by learning from data and making predictions Data Privacy Attacks • In data privacy attacks, an attacker is potentially able to infer the data set used to train the model, thereby potentially compromising the privacy of the data
- 34. Risk of Artificial Intelligence • Data poisoning is the contamination of data used to train the AI/ML system, negatively affecting its learning process or output Training Data Poisoning • Depending on the implementation and use case, the AI system could potentially evolve at varying degrees. Some forms of AI could generate complexities that may accrue, evolve or worsen over time Testing and Trust • As an emerging technology, the awareness of (and hype related to) AI and the lack of adequate understanding of the technology could potentially give rise to trust issues with AI systems Lack of Transparency • AI systems could potentially amplify risks relating to unfairly biased outcomes or discrimination Bias
- 35. Challenges in AI governance Data may lead to biases in the decisions of AI systems AI is quickly evolving on a technological level Output of AI is often difficult to understand Complex ML algorithms Difficulty in Interpreting algorithm output High risk of ethic washing AI produces unexpected results that are partly beyond the control of an organization Unpredictable ML decisions Ineffective Data control
- 36. Artificial Intelligence at Global Level The most prominent AI ethics guidelines are the recently released OECD Principles on AI (G20 2019) UNESCO is working on a possible ‘normative instrument’ on the AI The United Nations Interregional Crime and Justice Research Institute are in the process of opening a Centre for Artificial Intelligence and Robotics in The Hague, Netherlands. The Council of Europe has been active in AI, these developments are included in the next section on Europe The 40th International Conference of Data Protection & Privacy Commissioners (ICDPPC) in 2018 which saw the Declaration on Ethics and Data Protection in Artificial Intelligence released by delegates from various national data protection and privacy authorities The Declaration sets out six guiding principles and calls for ‘common governance principles on artificial intelligence’ to be established The ICDPPC has also set up a permanent working group on Ethics and Data Protection in Artificial Intelligence
- 37. AI Global Multistakeholder Initiatives Some multinational corporations have also released their own ethics statements Since many of these corporations originate in the US, they are included later in the section on the US There is one group that may be considered truly global, and multistakeholder in its membership, namely the Partnership on AI As mentioned, some UN agencies are among its members, as well as NGOs (such as Article 19), academic research institutes (such as the Australian National University 3Ai Centre), public sector agencies (including the BBC), and technology firms such as Amazon but also Chinese giant Baidu, the Partnership on AI has released its 8 ‘Tenets’ (Partnership on AI n.d.). The World Economic Forum, funded by its member corporations from around the world, has commenced various activities on AI, principally through its Centre for the Fourth Industrial Revolution in San Francisco (US) Part of this Centre’s work is to co-design and pilot policy and governance frameworks including for AI with governments and corporations. In 2019 the WEF released a White Paper on the topic of AI governance
- 38. Conclusion Governance is the activity to “make and enforce rules, and to deliver services” its actors involved in governance encompass individuals, citizens, organizations, and systems of organizations in public, private, and non-profit sectors Business AI governance could be seen as the framework of laws, customs, and procedures utilized to guarantee that the organization's AI technology upholds and furthers its aims and goals Artificial Intelligence is rapidly developing technologically, opening new commercial prospects and chances for digital innovation The performance of a company is significantly correlated with its governance, governance systems are crucial for reducing AI-related problems and increasing the potential of AI within enterprises
- 39. Recommendations AI governance businesses should imbibe ethics as a discipline that necessarily reduces the likelihood of unethical or biased outcomes in the AI field Companies should abide by the evolving number of rules on data and Machine Language models ensuring transparency, explainability, and accountability that can be achieved in AI systems Principles alone should not be expected to govern AI, considerations must also be given to the risks associated with rules and requirements to promote the business goals There should be adequate provision for relevant staff training to achieve their AI benefits There should be proper alignment with cultural norms to ensure that explicit rules and requirements achieve their full potential There should be explicit regulation necessary to interpret the complex output of Machine Language algorithms
- 40. Prof. Godwin Emmanuel Oyedokun Professor of Accounting and Financial Development Lead City University, Ibadan, Nigeria Principal Partner; Oyedokun Godwin Emmanuel & Co (Chartered Accountants, Tax Practitioners & Forensic Auditors)