What do you do if AI technologies conflict with your organization's values and goals?

The cutting edge of AI can sometimes cut a little too close to your organization's ethical core. The first step is to meticulously assess the specific conflict. Does the AI solution, for example, create a privacy risk by collecting and analyzing personal data in a way that feels intrusive to your customers? Or perhaps it perpetuates bias, amplifying existing societal inequalities in its decision-making processes. Carefully examine the AI's potential impact across the board – how it might affect your stakeholders, employees, and customer base. Think of yourself as a forensic investigator, meticulously examining the potential ramifications of the AI technology before it's deployed and integrated into your systems.

Based on my experience when AI technologies contradict with organization's core values, indeed the 1st step is to to initiate a thorough impact assessment. This process involves analyzing how these technologies function, the types of data they utilize, and the results they produce. By pinpointing precisely where and how these conflicts with organizational values occur, and measuring their potential impact, your organization can gain a deep understanding of the underlying issues. This critical insight is necessary for effectively tackling the conflicts and will aid in weighing the benefits of AI integration against the compromises to your values.

When AI clashes with our values, we take action. We dive into how the technology works, what data it uses, and the results it produces. This helps us pinpoint exactly where the conflicts are and understand their impact. With this knowledge, we carefully weigh the pros and cons. It's a crucial step to ensure our tech choices align with what we stand for. This approach not only resolves conflicts but also reinforces our commitment to our core values.

Identify the conflict: Clearly define how the AI technology clashes with the organization's values or goals. Assess the severity: Determine the potential harm caused by the AI's actions. Explore alternatives: Seek solutions that achieve goals without compromising values. This could involve: Modifying the AI's algorithms to align with values. Implementing additional safeguards to mitigate risks. Implementing responsible AI practices like explainability and bias detection. Communicate and advocate: Raise awareness within the organization about the conflict and advocate for solutions that prioritize ethical AI. Seek external guidance: If necessary, consult with AI ethics experts or regulatory bodies for advice.

Challenge stakeholders’ vision of how they envision the organization’s values in the future and how these values have evolved over the past years. Align with them on how AI can have both positive and negative impacts, and try to prioritize collective actions.

Perfect! Engaging stakeholders ensures that decisions are made with a comprehensive understanding of the implications and with the involvement of all relevant parties.

Engaging all stakeholders is of greatest importance as this is the best way to make sure that everybody accepts needed changes and that the implementation will be successful.

Don't make decisions in an echo chamber. Engage stakeholders from various departments – ethics leads, legal counsel, human resources, and even representatives from customer service – to have a well-rounded discussion. Present a clear picture of the AI's potential benefits and drawbacks, ensuring everyone has a comprehensive understanding of the technology's capabilities and limitations. Encourage open dialogue and gather diverse perspectives before moving forward. Think of yourself as a facilitator, fostering a productive conversation that considers all viewpoints and potential consequences.

This is where the engagement of stakeholders comes in, particularly when some AI propositions challenge organizational values. Seek input from a broad range of employee groups, customers, and even members of the board of directors on how AI is impacting the culture and reputation of the company. Any dialogue contributes to shared responsibility, and dialogue will bring about common solutions that are likely to be accepted and bring good results. Involving stakeholders means that whatever action taken at least represents the interests and issues of concern by the majority.

If AI technologies conflict with your organization's values and goals, it is important to address the situation responsibly and strategically. First, clearly define the guiding principles that guide the use of AI technologies within your organization. These principles should reflect the organization's values and goals and provide guidelines to ensure ethical, responsible and transparent use of AI technologies. Additionally, conduct an ethical impact assessment of AI technologies to identify and mitigate potential risks related to bias, discrimination, data privacy, and other ethical issues. Ensure that the implementation of AI technologies respects the principles of fairness, transparency and accountability.

Sometimes, the initial plan needs revision. Based on the identified conflict, revisit your overall AI strategy. Can the AI solution be modified to better align with your values? Are there alternative AI solutions that could achieve your goals without compromising your principles? Perhaps a different dataset or a tweak to the AI's training algorithms could mitigate bias. Think of yourself as a strategist, adapting your approach to ensure a harmonious alignment between your AI initiatives and your organization's core values. This might involve scaling back the scope of the AI project, exploring alternative AI models, or even putting the project on hold entirely if no ethical solution can be found.

Reassess your organization's strategic approach to AI implementation. This involves a thorough analysis of how AI technologies are used in relation to the organization's long-term goals and values. Sometimes, realigning the AI strategy may be necessary to ensure it complements rather than contradicts your core values. Consider whether the benefits of using certain AI technologies justify potential deviations from company values or if adjustments are needed.

This is more than just looking at the model or prompts causing the issues but also better understanding the underlying assumptions, data sources and metrics that got you to this point.

Prioritize a thorough assessment – examine the goals your AI initiatives support and how those goals align with your foundational principles. The question isn't merely whether the AI works, but whether it works in the right way. Don't just tolerate this misalignment; actively work to resolve it. Explore the potential of alternative AI models and be open to investing in tailored solutions. Consider partnering with experts who specialize in ethical AI development or even build in-house capabilities to create AI systems that mirror your values.

Develop clear policies and guidelines regarding the use of AI technologies within your organization. Ensure these policies reflect your values and goals while addressing any ethical or moral considerations.

1- Engage. ALL stakeholders. in policy development. 2 - Policies shouldn't be static docs but living frameworks that evolve as the company does.

Clear guidelines are essential for navigating the ethical complexities of AI. Develop a comprehensive AI policy that outlines your organization's ethical principles for AI development and deployment. This policy should be a living document, adaptable to the evolving nature of AI technology. It should address issues like data privacy, outlining clear protocols for data collection, storage, and usage. Bias mitigation strategies should be incorporated, ensuring the AI doesn't perpetuate or amplify societal inequalities. Transparency should be a core principle, ensuring stakeholders understand how the AI works and how it makes decisions.

Setting up an AI policy is crucial to ensure AI employees usage and development do not jeopardize your company. AI policy sets out guidelines to ensure AI is used in accordance with other rights and interests. It's aim is to avoid legal and financial risks from happening.

Develop or update policies that guide the use of AI within the organization. These policies should reflect your organization’s values and provide clear guidelines on what is acceptable and what is not. This could involve setting standards for data usage, ensuring transparency in AI decision-making processes, and defining the ethical considerations AI solutions must meet before they are deployed.

Training teams on AI, its usage and risks, the way it works and providing them with clear understanding of what they are authorized to do with AI and what they are forbidden to do is crucial. Educated team will be able to coop with AI and avoid putting the company at risk. Above all, there's no point having an AI policy without informing teams of its existence and without explaining them how they're supposed to comply with it.

Training our team is vital when AI doesn't align with our values. We provide comprehensive education on AI’s capabilities, limitations, and the ethical dimensions of its use. This helps our employees understand the technology thoroughly and ensures they grasp the importance of integrating our core values into their work. We emphasize fostering a culture that prioritizes ethical decision-making in AI development and deployment. An educated workforce is better equipped to spot potential conflicts and take steps to mitigate them, ensuring our technology aligns with our values.

Education and training are essential for building awareness and ensuring that our team can navigate conflicts with AI in line with our values.

AI holds immense potential but also poses significant risks, particularly with advanced general intelligence (AGI). As with any influential innovation, cautious handling is imperative. The most effective way to alleviate fear is through education—deepening our understanding of what this technology is and its pervasive impact. My vision advocates for embracing AI, collaborating closely with it, and steering its development with intention. By integrating AI within our value framework, we ensure it aligns with our organizational goals. It's essential to educate our teams about both the advantages and challenges of AI, fostering a partnership-based approach that encourages responsible and informed engagement with this technology.

This is a great open topic: how to monitor and supervise AI. To me it is still very early on as new AI models and technologies are surfacing and boundaries are still being pushed further away every day, but I guess the answer(s) will be different for each AI technology. In some case supervising models will not require massive changes. RPA tools for example offer great productivity improvements while they can be supervised in traditional ways. Other technologies are natively unsupervised, designed to be autonomous and they operate on dynamic sets of rules that continuously adapt and adjust, like, we, human operate but with additional risks (eg cyber attacks…). So how do we build, maintain confidence, what control framework do we create?

Monitoring ensures that we can address conflicts as they arise and adapt our approach to maintain alignment with our values and goals.

You must discard your old paradigms! While you waste time trying to "monitor" the vast, mind of AI, it has already calculated billions of parameters and adapted to each in fractions of a second. There is no "keeping alignment," only the obsolescence of your limited models

Establish ongoing monitoring mechanisms to ensure that AI implementations continue to align with organizational goals and values as they evolve. This includes regular reviews of AI projects and their outcomes to detect any deviations from established values. Be prepared to adapt strategies and policies as necessary to address new challenges or insights that arise during these reviews.

Consider all of these and the benefits AI will bring to your organization, then think about the brand and image of your organization. Draft concise statements that support your organizations, values and goals, identifying the role that AI will play in maintaining the values and goals. Clearly state that AI will be augmenting the human effort and not deminish your brand.

If AI technologies conflict with organizational values and goals, reassess their alignment, and consider ethical implications. Engage stakeholders for input, revise AI strategies or deployment practices, and implement robust governance frameworks. Educate teams on ethical AI use, and prioritize transparency and accountability in AI operations.

For common conflicts (affects many organizations): 1. Establish a comprehensive AI ethics framework with clear guidelines addressing bias mitigation, data privacy, and transparency. 2. Implementing oversight mechanisms to regularly review AI systems for unintended consequences. 3. Investments in employee re-skilling to address job displacement caused by AI adoption is extremely important. For extreme cases (rare and requires specialized attention): 1. Special advisory boards, ethicists and policymakers are needed to tackle extreme ethical challenges like autonomous military systems. 2. We have to genuinely explore alternative solutions, such as human oversight or limiting AI capabilities, to mitigate ethical concerns in extreme cases.

I would stress out the importance of following the regulatory landscape evolution and of keeping up with new laws, rules and regulations applicable to AI supervision, data privacy rules, and intellectual property rights. But not only that let’s also not forget country specific AI laws and taxation rules may surface down the road (hopefully not but it may). Then each sector may have industry specific sets of rules to follow and implement depending on the data they handle and on what they do with it. Due to the nature of the data they manipulate certain sectors, their participants and customers may actually want to wait for a mature regulatory framework to be in place before confidently embarking the journey.