AI in the Public and Private Sector
What are some opportunities challenges for governments in the use of AI in the public and or the private sector?
Blogger: Caroline Carlini
Artificial Intelligence (AI) is a quite recent phenomenon, full of progress perspectives, leading to great opportunities. However, it also poses multifaced legal ethical challenges both in the public and the private sectors, notably for governments. Rowena Rodrigues in her article « Legal and Human Rights issues of AI: Gaps, Challenges and Vulnerabilities.”, Published in the Journal of Responsible technology, in 2020, analyze how AI presents clear human rights issues.
The most known concern about AI, as an individual point of view at least, is the lack of accountability, transparency and contestability in the public sector.
The lack of algorithmic transparency leads to opaque decision-making processes and a lot of misunderstanding. It is almost impossible to ask for a justification concerning AI decisions because of a poor access to information. That is why it leads to a big lack of trust. Rowena Rodrigues point out that the solution to that could be algorithmic impact assessments as implemented by Canada’s government, but it still a new option that needs hindsight.
Concerning contestability, the problem comes from the fact that it is impossible to determine who is legally responsible for AI actions. It can be the developper, the government agencies/ the user or the AI itself. It appears crucial to have a legal framework on AI to give power to people or government that use this tool. The European Union implemented the GDPR (General Data Protection Regulation) Act that gives individuals rights to challenge and request a review of automated decision-making that significantly affects their rights or legitimate interests. It is a solution implemented but it still important to determine what is the AI’s criminal responsibility to enable potential prosecution.
This first concern is closely tied to intellectual property and privacy issues. Indeed, AI’s ability to create content also raises questions about ownership and rights, challenging existing intellectual property frameworks. We can wonder, who owns the data on which the AI is based ? Who owns what AI machines produce ? Rowena Rodrigues adds, that often, AI relies on vast amounts of personal data, heightening risks related to privacy breaches and unauthorized data usage. For instance, some artists, writers or musicians have prosecuted OpenIA for using their work without any consent, compensation, or attribution, in violation of U.S. copyright law. According to Christina Pazzanese in her The Harvard Gazette’s article published in 2023, « OpenAI has said the books are used only to spur innovation, not to create new works, and that that practice is lawful under the “fair use” provision of copyright law », when the authors and artists talk about « derivative works ». The article also tackles the main point that we point out before : who is legally responsible on that matter ?
As another example, deepfakes illustrate also this matter. This term refers to images, videos or audio generated by AI and that may depict real or non-existent people. Recently, French president Macron reacted to one of this edited deepfakes videos of him, quite laughing rather than worrying. All seems to be about saving face in front of the people while fighting against this big and powerful invention. Meanwhile, the UK has passed a law that protects computer-generated literary, dramatic, musical or artistic works. The rights are vested in the creator of the AI.
In the same vein, AI is also at the origin of discrimination and injustice perpetuation. IA is trained on historical data that can be dangerously biased. Job interview selections or faced recognition are examples among others that can be based on algorithms. The National Institute of Standards and Technology (NIST) led a demographic study on facial recognition software to assess their ability to perform different tasks. The study noticed two types of errors : false positives and false negatives. A false positive means that the software wrongly considered photos of two different individuals to show the same person, while a false negative means the software failed to match two photos that, in fact, do show the same person. Discriminatory errors can be easily made on physical appearance and lead to very problematic situations. According to Innocence Project, in 2020, Robert Williams a black Detroit inhabitants spent 30 hours in custody because police facial recognition software incorrectly assumed that a thief’s face, captured on a security camera while stealing from a jewelry store, was identical to the photo of his driver’s license.
In 2017, the European Parliament made a report on the topic calling the European Commission, Member States and data protection authorities "to identify and take all possible measures to minimise algorithmic discrimination and prejudice, and to develop a strong and common ethical framework for the transparent processing of personal data and automated decision-making that can guiding the use of data and the continued application of Union law ».
Then, AI is also challenging the working sphere. It has an impact on workers in term of requirements, working time, remuneration, job displacement… Nowadays, in every public platforms we can find chatbots. This improvement is in fact a real threat for workers specialized in that service. To make up for that, workers have to be trained in other fields, to follow up-skilling training to find new job. In any case, it raises economic and social human rights issues (poverty, displacement, despair). Another examples is the use of AI for voices dubbing. In October 2024 in Quebec, thousands of voice dubbing artists expressed concerns about the use of AI to reproduce or resurrect human voices, potentially threatening their profession. Sometimes, AI can even simulate these artists own voice. They claim legal frameworks that allow them to challenge the unauthorized use of their work by AI systems. This challenge is of course related to privacy and legal responsibility as well.
Naturally, cybersecurity is a huge challenge in the use of AI. While being pivotal in safeguarding computer networks against cyber threats such as malware, phishing and unauthorized access, AI systems are often not enough protected against cybercrimes. According to S.S.A. Bokhari and S. Myeong’s article, in IEEE Xplore, it has been alleged that cyber-attacks have become more prevalent in the industrial sector, resulting in serious infrastructure damage and significant monetary loss. It seems to be due to the growing reliance on online technologies that enable the storage of personal and economic data. The introduction of E-governance that relies on ICTs, also jeopardize the citizens that frequently use insecure Wi-Fi networks to check their email messages, e-banking, and other digital services, uncovering themselves to cybercrimes including hacking, denials of service, and cracking. The article basically explains how smart cities, that is to say cities that rely on smart government, smart economy, smart healthcare, smart security, smart public services etc, are highly vulnerable to cybercrimes even if AI tools appear as the best way to thwart these types of attack. Misinformation and deepfakes were heavily used during general elections in France, US and even England, particularly by Russia. It has leaded to rigged elections not representing the will of the people and to the empowerment of extreme-right leaders that follow Putin’s ideas. Interference is one of these cyberthreats.
S.S.A. Bokhari and S. Myeong focus also on stakeholder’s role to participate actively, to enhance both the acceptance and success of AI initiatives. It reinforces the idea that AI adoption is not just a technical issue but also a social and political one.
To conclude, AI a new revolutionary tool that may help and ease life from an individual and government’s perspective but at the same time that need to be urgently regulated regarding all these challenges and threats on the public order.
Questions:
1. Have you ever felt scared regarding AI progresses? Explain why.
2. As a future job seeker, are you afraid of the consequences of AI on job availability? And, as a stakeholder what would you offer to ensure AI implementation aligns with national security priorities?
Resources:
Christina Pazzanese, «Is ChatGPT more foul than fair?», The Harvard Gazette, 2023
https://news.harvard.edu/gazette/story/2023/09/key-issues-in-writers-case-against-openai-explained/
NIST Study Evaluates Effects of Race, Age, Sex on Face Recognition Software, "Demographics study on face recognition algorithms could help improve future tools », 2019.
Alyxaundria Sanford, «Artificial Intelligence Is Putting Innocent People at Risk of Being Incarcerated», Innocence Project, 2024
S.S.A. Bokhari and S. Myeong, « The Influence of Artificial Intelligence on E-Governance and Cybersecurity in Smart Cities: A Stakeholder’s Perspective », IEEE Xplore, 2023https://ieeexplore.ieee.org/document/10177170
Helène Jouan,“Les artistes québécois s’interrogent face aux progrès de l’IA.”, Le Monde, 2024https://www.lemonde.fr/culture/article/2024/10/08/les-artistes-quebecois-s-interrogent-face-aux-progres-de-l-ia_6346694_3246.html
The blogger provided a thorough review of AI hazards and legal problems supported by real-world examples. However the blog gave more information about the dangers of AI while not fully exploring its benefits. The inclusion of policy recommendations would have been valuable.
ReplyDelete1. Have you ever felt scared regarding AI progresses? Explain why.
Yes, I have felt scared regarding AI progresses. In my opinion, while AI brings many benefits, it also poses significant risk. One major concern is its potential to threaten national security. Additionally, the fear that AI could replace human jobs, leading to economic challenges. Also, AI’s growing influence may contribute to decline in trust within the society.
2. As a future job seeker, are you afraid of the consequences of AI on job availability? And, as a stakeholder what would you offer to ensure AI implementation aligns with national security priorities?
As a future job seeker, yes, I am afraid of the consequences of AI on job availability. Routine-based and repetitive jobs are high risk of automation, especially in sectors like customer service, data entry and manufacturing. As result, job seekers need to develop adaptability and continuous learning, as this is crucial for them to remain relevant in an AI-driven job market.
As a stakeholders, ensuring AI implementation aligns with national security would require a multi-faceted approach. This includes regulating AI-powered misinformation by establishing laws against AI-generated fake content to prevent electoral interference, as misinformation has been increasingly used to manipulate public opinion. Also by investing in cybersecurity is important to strengthen AI-driven security frameworks that detect and prevent cyber threat.
Trisha Bowes Green, I believe one major sector that will benefit greatly from AI is the health sector. If you should delve into the 4th and 5th Industrial Revolutions, you will be amazed to see the things that AI is doing. One that stood out was the Robot that did open-heart surgery just by making a simple incision under the rib cage. The patient was out and about within days, doing exercise! That was amazing! Both public and private sectors will now have to retrain, retool and re-skill employees as some job functions will become obsolete due to AI. My answers to the questions are as follows:
ReplyDeleteAI has the potential to revolutionize various industries by enhancing productivity, creativity, and decision-making. However, it also poses challenges such as privacy, security, and potential unforeseen consequences. AI systems, particularly those using machine learning, may reinforce biases in training data, leading to unjust outcomes in critical areas like recruitment, law enforcement, and loan approvals. The lack of transparency in decision-making processes and uncertainty surrounding accountability for errors in AI systems are troubling issues.
The rise of AI-driven automation presents both opportunities and challenges for job seekers. While AI can free employees from mundane tasks, it can also lead to job displacement in sectors like manufacturing, customer service, and transport. Transitioning the workforce to new job opportunities while mitigating unemployment and economic inequality requires robust governmental interventions, investments in education and training programs, and strengthening social safety nets.
To ensure AI adoption adheres to security interests, strategic measures such as establishing AI Ethics and Safety Guidelines, public-private partnerships, and education and public awareness are necessary. Effectively addressing these challenges through legislation, ethical standards, and rigorous security protocols will be crucial for harnessing AI's benefits while safeguarding societal interests.
This comment has been removed by the author.
ReplyDelete1. Have you ever felt scared regarding AI progresses? Explain why.
ReplyDeleteAbsolutely, but not in the dystopian sense that is talked about a lot. The basis of my concern revolves around the often hidden qualities of decision-making processes in artificial intelligence systems. As Rodrigues (2020) explains, one of the main threats posed by artificial intelligence is a failure of transparency and explainability, especially when algorithms are used in decision-making settings, i.e., policing or credit checks. It's not just that you will never know how a particular decision was reached. This is not merely a technical issue and it is also a democratic issue. When you simply delegate decision-making from people to machines in a way that is not publicly accountable, you end up degrading public accountability and trust in the democratic process.
2. As a future job seeker, are you afraid of the consequences of AI on job availability? And, as a stakeholder what would you offer to ensure AI implementation aligns with national security priorities?
I see myself as a future job seeker not caught in job loss, but skills will have changed in demand. UNDESA's article (2024) describes work that has the potential to enhance benefits to experience and efficiencies of performance in operations of many public services automated, still depending on a foundation of digital fluency and adaptable learning. The threat is not job loss itself, but our own ability to be ready to change. In the role of stakeholder, acting with AI at the level of national security sees the need to move beyond just an understanding of cybersecurity to rationale for digital sovereignty and ethical resilience. I would be supportive of AI regulatory sandboxes such as is being proposed in the new governance system in the possible use of the government running AI systems before use is refined in identified testing spaces of policy development. Careja & Tapus (2023) considered we can use blockchain for digital identity verification to be able to enforce trust in a spectrum of digital interactions while emphasizing the user is still protected from disinformation and unreasonable use of the data. The encouraging aspects of this process exist in rethinking adaptation and governance.
To answer question 1, the progress of AI worries me in all areas. As Caroline explained, AI is not yet regulated and could replace many people, especially in low-skilled jobs. I wanted to follow up on the argument about film dubbing and art in general. In France, lastJanuary, voice actors launched a petition to support the French version of films against AI. A famous French actor, Alain Dorval, who dubbed Silvester Stallone recently died, and his daughter (who is also a member of the current government) gave her permission to use AI to reproduce her father's voice. The result being very "bland", she explained that she had not authorized the release of the trailer re-dubbed with AI. So we could wonder if AI in art will become democratized, especially given that certain members of the government have encouraged this practice, and like Caroline said, even the French president seems to like it (especially he made that video at the same time that this debate took place). AI doesn't create anything but uses other people's data to meet a demand. Today, many artists are being competed by pseudo-artists who don't produce anything but use AI to produce bad content stolen from other artists. Thus, people who hire artists, such as companies or individuals, are increasingly abandoning artists because they use AI or hire people who know how to use AI because they are cheaper. On the internet, images generated by AI are found on the front page and on all media and websites (press articles, youtube...). Concerning the field of dubbing, actors are currently asking for the protection of their biometric data, such as their voices or faces. Even though legislation is beginning to be adopted in Europe, companies based outside the continent are ignoring European legislation to use this data. So AI is already stealing actors' biometric data; what's stopping it from stealing from "normal" citizens? Although it seems unlikely at the moment, it is conceivable that images from street surveillance systems could be sold to private companies and that the faces of passers-by could be "monetized" in the future.
ReplyDelete1. No, I am not fearful of how AI progresses, but rather how some people are exploiting its capabilities to commit unlawful and unethical acts. Like many technological advancements, AI is developed with the intention of enhancing human life, making processes more efficient, information more accessible, and services more personalized. For example, just as the invention of the telephone revolutionized real-time communication across vast distances, now that technology is also used to scam. AI similarly offers groundbreaking potential across both public and private sectors. When used ethically, AI can drive significant improvements in productivity, streamline services, and reduce wait times by handling multiple queries simultaneously. For instance, instead of having to wait in long queues to speak with a human representative, AI chatbots can address multiple clients at once, offering a faster and more efficient customer service experience. However, it should not be viewed as a replacement for human interaction, but rather as a complementary communication channel that supports human workers. That said, there are very real concerns associated with AI, especially because it is programmable and data-driven. The integrity of AI outputs depends heavily on the data it is trained on. This raises issues of bias, misinformation, and overreliance. Many people increasingly turn to AI as a primary source of information due to its ability to summarize complex material quickly. However, this ease of access may lead to blind trust in its responses, without verifying the credibility or accuracy of the information. In this way, individuals may unintentionally internalize incorrect or misleading content. One of the more alarming developments in AI misuse involves deepfakes (as you mentioned). Deepfakes can impersonate voices, mimic facial expressions, and even replicate entire appearances of real people with astonishing accuracy. While this technology has some legitimate applications in entertainment and education, it has also become a tool for deception. This can be seen as a more advanced and dangerous evolution of “catfishing,” whereby complete strangers in any part of the world can masquerade as a completely different person online or using telephony to commit fraudulent activities.
ReplyDelete2. As a future job seeker, I acknowledge that AI will have an impact on job availability, particularly in areas such as administrative support, data entry and customer service However, I am not afraid of the consequences of AI. Rather than viewing AI as a threat to employment, I believe it should be seen as a call to action—a push for individuals to upskill and reskill in order to remain relevant and competitive in a rapidly evolving job market.Therefore, It’s crucial that both governments and private sector stakeholders invest in robust reskilling and upskilling initiatives to help workers transition into these new roles. Though there may be a replacement of certain tasks, it also creates new roles that require human oversight, creativity, emotional intelligence, ethical reasoning, and decision-making areas where machines fall short as such I believe that human beings remain essential to the entire AI ecosystem. From a stakeholder’s perspective, ensuring AI implementation aligns with national security priorities requires a multi-pronged approach Strengthen Cybersecurity Infrastructure:
1. A well-informed public is the first line of defense- Stakeholders should invest in public education campaigns to teach people how to spot fake news, keep their personal info safe, and use AI tools ethically.
2. A collaboration with other countries - Countries should work together to establish global norms, share intelligence on AI threats, and coordinate responses to cyberattacks and disinformation campaigns.
3. Upgrade Cybersecurity: Governments need to invest in AI-powered defenses that can spot and fight back against new threats like phishing scams, AI-made scams, and fake videos.