{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Introduction\nWhat is Ethics? It is the branch of philosophy that deals with moral principles. It teaches us what is right and what is not. Although there are many grey areas regarding right and wrong, most of them are straightforward. Murder - wrong. Robbery - wrong. Philanthropy - right. However, it is not the same when it comes to dealing with new technology.","metadata":{}},{"cell_type":"markdown","source":"# Who?\nOne of the most important problems that we face in this area is that almost all Ethics committee members are too outdated to fully understand and comprehend the technology to its fullest extent. This makes it incredibly hard to make any progress in this space.","metadata":{}},{"cell_type":"markdown","source":"The people who have enough knowledge of this technology and can make a difference are the big tech giants such as Google, Facebook, Microsoft, Amazon, and IBM. They have also made strides by starting a nonprofit back in 2016 called \"Partnership on Artificial Intelligence to Benefit People and Society\" to help academics and ethics experts. However, there is still a question at large. Can they be trusted? Can these big tech companies be trusted to put the people first and the money second?","metadata":{}},{"cell_type":"markdown","source":"# What?\nThen there's this debate. The importance of learning to ride a horse is way more valuable in the 1800s compared to the present. Driving an automobile is a basic necessity now. NASA does not hire human calculators anymore, because they have computers to do it for them. Now they hire people who can effectively operate these advanced computers. Therefore, it is understood that along with the evolution of science, engineering, and technology, the skills that an average person needs to learn to get by in life also change. So, one might argue that the introduction of generative AI applications like ChatGPT can have the same effect.","metadata":{}},{"cell_type":"markdown","source":"Facebook was launched in 2004. Twitter, in 2006. Instagram, in 2010. Many other social media platforms have made their debut after that. It has been nearly 20 years and there are still many ethical challenges that have not been remedied yet. Even the ones that made it to the ethics books are rarely enforced effectively. So, there is a lag. A lag between technological advancement and ethics enforcement that protects the vulnerable from these technologies. This lag has always existed. Between the invention of consumer automobiles and the regulations that made them safer, the invention of the world wide web and the regulations that followed protecting the people using them, and even now the status quo remains the same with the advancements in the AI industry. However, the time lag seems to get longer and longer every time. And this time, this lag can potentially create many irreversible damages to society as we know it now.","metadata":{}},{"cell_type":"markdown","source":"Can we do something different this time to mitigate those damages? How do we close the knowledge gap between the people who run the tech companies and the people who regulate them? Should we take a more aggressive approach? And risk getting better?","metadata":{}},{"cell_type":"markdown","source":"> Endowments are an insidious way for wealthy benefactors to exert influence on universities, guide their research including policy proposals, and it is not realistic to expect donors to fund any academic initiatives to reform a system that directly or indirectly benefits them. (Mia Dand, CEO of Lighthouse3)","metadata":{}},{"cell_type":"markdown","source":"The rich even influence the most elite of universities. It is very unlikely that they would be supportive of AI Ethics researchers working on strict AI regulatory research which would cost these rich their riches.","metadata":{}},{"cell_type":"markdown","source":"According to Jobin (2019), there is only one AI Ethical Guideline (published by the Royal College of Physicians, United Kingdom) among the 84 major guidelines they analyzed even mentioning the risk of technological progress outpacing the regulatory measures. Not a single guideline talks about protecting children from AI applications as well.","metadata":{}},{"cell_type":"markdown","source":"> When a decision-making system's usefulness, accuracy, or underlying logic is misaligned with users' expectations, it can induce misplaced trust resulting in harm. (AI & Machine Learning, Center of Democracy & Technology)","metadata":{}},{"cell_type":"markdown","source":"# Analysis\nInspired by the research of Jobin (2019), this attempt is a follow-up research to see how the Artificial Intelligence Ethical Principles and Guidelines published by multiple parties have faired after that. The search identified 33 guidelines published across the world wide web by Universities, Private companies, Governments, etc. between January 2020 to June 2023. In order to maintain the search unbiased, all the search was conducted in a private browsing window with all personal accounts logged out and cookies and browsing history deleted.","metadata":{}},{"cell_type":"markdown","source":"![Table 1](https://storage.googleapis.com/kagglesdsdata/datasets/3522761/6143217/Screenshot%202023-07-15%2012.49.19%20PM.png?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20230716%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20230716T102706Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=5c75b8f79aab91fe5ecc08a8de9ebd84a657a902fd2102ea397e661e61734c0bd803b7b56643dcf7fc0b21bd8c874e467738ce0ddd8a6d92d657cb1f685bf339d71f5e6a03c379df3aeaab2481dca663cf6dbd345e9c46cefcc02f33dc4cad52004b1cd10842bbc0ad847fa475666ad08c3f8215bae3e11fc1f648834cf6db7274b12c1ff3c05b78462d243691fedf764af5c187563c6ab45c23595cdd2f9eeb92147025f678aed00ad1fa91001f3fa2b8e7dcb0a80c409a054c06356e4390db3aacbfd22644ab5aab049fc33c40615bfd5db5a4fb89485c42129e492748eb44fdab1bd2318aca2fff2b078c65a1c6f7bde6dcd83c36c5c598334b42650b47f9)","metadata":{}},{"cell_type":"markdown","source":"Jobin's research (2019) talks about 11 ethical principles that are very important to Artificial Intelligence. This research uses all of those 11 principles - Transparency, Justice & Fairness, Non-maleficence, Privacy, Responsibility, Beneficence, Freedom & Autonomy, Trust, Sustainability, Dignity, and Solidarity. One more very specific case scenario is also added to the list making it 12 - Exposure to Children.","metadata":{}},{"cell_type":"markdown","source":"Almost no one talks about how the new AI is harming children. Now just AI, any new technology for that matter. Social media such as Instagram and Tiktok are setting unrealistic body standards for school kids. Now the new generative AI tech is doing their homework for them. AI tools do help professionals be more productive and efficient. But that is all they should be - tools, for professionals. Any 13-year-old kid with a Google Mail account can access ChatGPT. And the only reason companies even have an age restriction of 13 is that the Children Online Privacy Protection Act (COPPA) says so. There should be more strict age restrictions, for the right reasons. And these restrictions need to be enforced properly rather than companies using that as a way of absolving them from any legal liability.","metadata":{}},{"cell_type":"markdown","source":"![Table 2](https://storage.googleapis.com/kagglesdsdata/datasets/3522761/6143217/Screenshot%202023-07-15%2012.15.39%20PM.png?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20230716%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20230716T102701Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=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)","metadata":{}},{"cell_type":"markdown","source":"The data analysis utilizes  chatpdf's API which is powered by OpenAI's GPT 3.5 Large Language Model to find out which principles are mentioned in which guidelines. It was surprising to learn that only 2 (AI Ethics - Samsung SDS & AI Ethics - Kakao) out of the 33 mentioned taking caution while exposing AI to children. This needs to change because it is important, to protect the children from over-exposure to AI technology. The hope is that this research would pave the way for that, like how Jobin's (2019) research impacted this community for the better. Just to elaborate on that point, here's an example. Only 55 percent of the AI Ethical Guidelines mentioned \"privacy\" before 2019, according to Jobin. In the last four years, that has increased to more than 85 percent. 29 out of the 33 guidelines analyzed mentioned the importance of privacy in Artificial Intelligence.","metadata":{}},{"cell_type":"markdown","source":"![Chart 1](https://storage.googleapis.com/kagglesdsdata/datasets/3522761/6143217/chart%20%285%29.png?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20230716%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20230716T103111Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=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)","metadata":{}},{"cell_type":"markdown","source":"The guidelines analyzed are chosen across multiple sectors, across different countries, and across different years to maintain diversity. However, there are limitations to this research. The majority of the guidelines were published in the United States and Europe. There is very little to no representation of Asian, African, and South American countries. According to SimilarWeb, more than 6 percent of ChatGPT's users are from Brazil and Columbia. However, this is not reflected in the number of AI Ethics Guidelines published in these countries. The language barrier was another challenge that further research on this topic should overcome.","metadata":{}},{"cell_type":"markdown","source":"![Chart 2](https://storage.googleapis.com/kagglesdsdata/datasets/3522761/6143217/Sector.png?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20230716%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20230716T103255Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=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)![Chart 3](https://storage.googleapis.com/kagglesdsdata/datasets/3522761/6143217/Country.png?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20230716%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20230716T102620Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=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)![Chart 4](https://storage.googleapis.com/kagglesdsdata/datasets/3522761/6143217/Year.png?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20230716%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20230716T103257Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=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)","metadata":{}},{"cell_type":"markdown","source":"# Conclusion\nOverall, the question still remains at large. Who decides what? Who should take up more responsibility? What AI Ethics principles should we prioritize? To that, my answer is - Everyone and Everything. We should all step up. Try to achieve more convergence in the AI Ethics principles. Starting a dialogue and keeping on talking is our only hope. And that is what I am doing here.","metadata":{}},{"cell_type":"markdown","source":"# References\n\nRecommendation on the Ethics of Artificial Intelligence, UNESCO (2021) \nhttps://unesdoc.unesco.org/ark:/48223/pf0000381137\n\nEthics of Artificial Intelligence - Our AI Ethics Guidelines, Bayerischer Rundfunk (2020) \nhttps://www.br.de/extra/ai-automation-lab-english/ai-ethics100.html\n\nEthical Guidelines on the use of Artificial Intelligence, European Commission (2022) \nhttps://op.europa.eu/en/publication-detail/-/publication/d81a0d54-5348-11ed-92ed-01aa75ed71a1/language-en\n\nEthics and Governance of Artificial Intelligence for Health, WHO (2021) \nhttps://www.who.int/publications/i/item/9789240029200\n\nArtificial Intelligence Principles of Ethics and Framework for the Intelligence Community, National Intelligence (2020) \nhttps://www.intelligence.gov/principles-of-artificial-intelligence-ethics-for-the-intelligence-community\n\nFrom Principles to Practice - How can we make AI Ethics Measurable, Bertelsmann Stiftung (2020) \nhttps://www.bertelsmann-stiftung.de/en/publications/publication/did/from-principles-to-practice-wie-wir-ki-ethik-messbar-machen-koennen\n\nDOD Adopts Ethical Principles for Artificial Intelligence, Department of Defense (2020) \nhttps://www.defense.gov/News/Releases/Release/Article/2091996/dod-adopts-ethical-principles-for-artificial-intelligence/\n\nA Practical Guide to Building Ethical AI, Harvard Business Review (2020) \nhttps://hbr.org/2020/10/a-practical-guide-to-building-ethical-ai\n\nAI Ethics - IBM, IBM (2023) \nhttps://www.ibm.com/impact/ai-ethics\n\nBosch Code of Ethics for AI, Bosch (2020) \nhttps://www.bosch.com/stories/ethical-guidelines-for-artificial-intelligence/\n\nOn the Governance of Artificial Intelligence through Ethics Guidelines, Lund University (2020) \nhttps://www.cambridge.org/core/services/aop-cambridge-core/content/view/992BD33CA7CBBE83E2FBBF6B0179896C/S2052901520000194a/on-the-governance-of-artificial-intelligence-through-ethics-guidelines.pdf\n\nGovernance Guidelines for Implementation of AI Principles, Ministry of Eco0mic, Trade and Industry (2022) \nhttps://www.meti.go.jp/shingikai/mo0_info_service/ai_shakai_jisso/pdf/20220128_2.pdf\n\nPrinciples for the Ethical Use of AI in the UN System, United Nations System (2022) \nhttps://unsceb.org/sites/default/files/2022-09/Principles%20for%20the%20Ethical%20Use%20of%20AI%20in%20the%20UN%20System_1.pdf\n\nPosition Paper of the People's Republic of China on Strengthening Ethical Governance of Artificial Intelligence (AI), Ministry of Foreign Affairs of the People's Republic of China (2022) \nhttps://www.fmprc.gov.cn/eng/wjdt_665385/wjzcs/202211/t20221117_10976730.html\n\nOn a Risk-Based Assessment Approach to AI Ethics Governance, Institue of Ethics for Artificial Intelligence (2022) \nhttps://www.ieai.sot.tum.de/wp-content/uploads/2022/06/IEAI-White-Paper-on-Risk-Management-Approach_2022-FINAL.pdf\n\nAn Ethics Label for AI - University of Stuttgart, University of Stuttgart (2021) \nhttps://www.uni-stuttgart.de/en/research/forschung-leben/1-2021/ethics-label/\n\nLofty Principles, Conflicting Interests - AI ethics and governance in China, Merics (2021) \nhttps://merics.org/sites/default/files/2023-02/MERICSChinaMonitor69AIEthics4.pdf\n\nSAP Global AI Ethics Policy, SAP (2022) \nhttps://www.sap.com/germany/products/artificial-intelligence/ai-ethics.html?pdf-asset=a8431b91-117e-0010-bca6-c68f7e60039b&page=1\n\nGoverning AI - A Blueprint for the Future, Microsoft (2023) \nhttps://query.prod.cms.rt.microsoft.com/cms/api/am/binary/RW14Gtw\n\nAI Ethics - How To Act Ethically in an Automated World, Ultimate (2022) \nhttps://www.ultimate.ai/blog/collaborations/ai-ethics\n\nMandatory Ethical Principles for the use of AI, NSW (2020) \nhttps://www.digital.nsw.gov.au/policy/artificial-intelligence/artificial-intelligence-ethics-policy/mandatory-ethical-principles\n\nSGModelAIGovFramework2, Personal Data Protection Commission (2023) \nhttps://www.pdpc.gov.sg/-/media/Files/PDPC/PDF-Files/Resource-for-Organisation/AI/SGModelAIGovFramework2.pdf\n\nIFPMA Artificial Intelligence Ethics Principles, IFPMA (2023) \nhttps://www.ifpma.org/wp-content/uploads/2023/02/IFPMA-Artificial-Intelligence-Ethics-Principles-toolkit.pdf\n\nAI Ethics Guidelines That Organizations Should Consider, Forbes (2023) \nhttps://www.forbes.com/sites/forbestechcouncil/2023/06/15/ai-ethics-guidelines-that-organizations-should-consider/?sh=512f2fc17f7c\n\nAI Ethics and Governance, Fujitsu (2022) \nhttps://www.fujitsu.com/global/about/research/tech0logy/ai/aiethics/\n\nAI ethics - The moral and ethical implications of using artificial intelligence, HP (2022) \nhttps://community.hpe.com/t5/tech-insights/ai-ethics-the-moral-and-ethical-implications-of-using-artificial/ba-p/7163353\n\nEthical AI, C3 (2020) \nhttps://c3.ai/glossary/artificial-intelligence/ethical-ai/\n\nTrustworthy AI in Aotearoa, AI Forum (2020) \nhttps://data.govt.nz/assets/data-ethics/algorithm/Trustworthy-AI-in-Aotearoa-March-2020.pdf\n\nORF_IssueBrief_589_AI-Ethics, Observer Research Foundation (2022) \nhttps://www.orfonline.org/wp-content/uploads/2022/11/ORF_IssueBrief_589_AI-Ethics.pdf\n\nAI Ethics - Samsung SDS, Samsung SDS (2022) \nhttps://www.samsungsds.com/en/digital_responsibility/ai_ethics.html\n\nAI Ethics _ Kakao, Kakao (2022) \nhttps://www.kakaocorp.com/page/responsible/detail/algorithm?lang=ENG&tab=all\n\nAI Ethics for Solution Developers, IEEE (2022) \nhttps://engagestandards.ieee.org/rs/211-FYL-955/images/AI%20Ethics%20for%20Solution%20Developers?_gl=1*j18tvf*_ga*ODUyMjgyOTk2LjE2ODkyNTU2MDA.*_ga_XDL2ME6570*MTY4OTI2MTIyMC4yLjEuMTY4OTI2MzkzMC4wLjAuMA..\n\nFive guiding principles for responsible use of AI in healthcare and healthy living - Blog _ Philips, Philips (2020) \nhttps://www.philips.com/a-w/about/news/archive/blogs/in0vation-matters/2020/20200121-five-guiding-principles-for-responsible-use-of-ai-in-healthcare-and-healthy-living.html\n\nArtificial Intelligence: The Global Landscape of Ethics Guidelines (2019)\nhttps://arxiv.org/pdf/1906.11668.pdf","metadata":{}}]}