Last updated on 2024年6月30日

How can you deploy responsible AI?

由人工智能和领英社区提供技术支持

Artificial intelligence (AI) can offer many benefits to society, but it also poses significant challenges and risks. How can you deploy AI in a way that respects human values, protects privacy, and promotes fairness and accountability? In this article, we will explore some key principles and practices for responsible AI, and how you can apply them to your own projects.

此文章中的业界达人

由社区从 12 条内容中精选。了解更多

Ripal Shah

Risk Manager - Banking & Financial Services: Expert ? Risk Transformation I Risk Frameworks I GRC Systems I Risk Culture
Ilia Amirmostofian

Top LinkedIn Voice in Automotive Engineering ?? Driving Sustainability & EV Innovations | Passionate about AI & Tech
Zaki E.

Data & AI Leader at Electronic Arts (EA) ?? | Designing Scale & Innovation ??

1 Define your purpose

Before you start developing or using AI, you need to have a clear and explicit purpose for it. What problem are you trying to solve, and how does AI fit into your solution? What are the expected outcomes and impacts of your AI, and who are the stakeholders involved? How will you measure and evaluate your AI's performance and value? By defining your purpose, you can align your AI with your mission, vision, and values, and avoid unintended or harmful consequences.

添加您的观点

Ripal Shah

Risk Manager - Banking & Financial Services: Expert ? Risk Transformation I Risk Frameworks I GRC Systems I Risk Culture
举报内容
AI works differently to traditional IT systems and has a fundamentally different risk profile. Organizations need to change their traditional business case templates to incorporates measurements and evaluation of AI's performance and value as well as unintended consequences and risk indicators to monitor them.

已翻译

赞
Ilia Amirmostofian

Top LinkedIn Voice in Automotive Engineering ?? Driving Sustainability & EV Innovations | Passionate about AI & Tech
举报内容
Deploying responsible AI involves several key steps. Start with establishing clear ethical guidelines that prioritize fairness, accountability, and transparency. Use diverse datasets to train models, regularly auditing for biases. Maintain transparency by documenting and explaining AI processes to stakeholders. Implement human oversight, especially in critical areas, ensuring humans can intervene when necessary. Continuously monitor AI performance and compliance, updating systems as needed. Engage with a wide range of stakeholders for feedback. Finally, stay informed about and comply with relevant regulations. These steps foster trust and maximize AI's positive impact. This is what I try to do in the personal projects that I'm engaged in.

已翻译

赞

加载更多内容

2 Assess your risks

AI is not a neutral or flawless technology. It can introduce or amplify biases, errors, vulnerabilities, and harms, depending on how it is designed, trained, tested, and deployed. Therefore, you need to assess the potential risks of your AI, and how they affect different groups and domains. You can use frameworks and tools such as AI impact assessments, risk matrices, or ethical checklists to identify and prioritize the sources and levels of risk, and to mitigate them accordingly.

添加您的观点

Ripal Shah

Risk Manager - Banking & Financial Services: Expert ? Risk Transformation I Risk Frameworks I GRC Systems I Risk Culture
举报内容
Ensure that risks are not viewed in silo. Benefits and risks need to be viewed together. Quantification of risks and benefits would be helpful to senior leaders and decision makers.

已翻译

赞
Riccardo Zanardelli

Head of Data Science @Beretta | PhD Student in Statistics and Data Science | Engineer | Only personal views here
举报内容
Make clear what you are responsible for, by assessing the most critical risk for your users. Share with you user base these risks, and make sure that they know how to weigh your predictions in their decisions. Make sure to collect their feedback... sometimes your definition of risk and responsibility might not overlap with your users' one.

已翻译

赞

加载更多内容

3 Engage your stakeholders

AI affects not only the developers and users, but also the wider society and environment. Therefore, you need to engage your stakeholders throughout the AI lifecycle, and solicit their feedback, input, and consent. You can use methods such as surveys, interviews, focus groups, or co-design workshops to understand the needs, expectations, and concerns of your stakeholders, and to involve them in the decision-making and governance of your AI. You can also communicate your AI's purpose, capabilities, and limitations to your stakeholders in a transparent and accessible way.

添加您的观点

Florian Baptiste

?? Empowering businesses to create impactful digital products in AI, Apps, and AdTech. Interested? Let's connect and explore opportunities together!
举报内容
Ensure that the aims of your endeavours are apparent to everyone involved. It's essential to set expectations with stakeholders about your intentions, particularly detailing what you can or cannot undertake, in order to achieve unified agreement.

已翻译

赞

加载更多内容

4 Implement your standards

To deploy responsible AI, you need to adhere to the relevant standards, guidelines, and regulations that apply to your context and domain. These can include technical standards for data quality, security, and interoperability, as well as ethical principles for human dignity, autonomy, justice, and welfare. You can use frameworks and tools such as AI ethics codes, certification schemes, or audit systems to ensure that your AI complies with the standards, and to demonstrate your accountability and trustworthiness.

添加您的观点

加载更多内容

5 Monitor your outcomes

AI is not a static or fixed technology. It can evolve and adapt over time, and produce unexpected or undesirable outcomes. Therefore, you need to monitor your AI's performance and impact on an ongoing basis, and update or adjust it as needed. You can use methods such as data analysis, feedback loops, or evaluation metrics to track and measure your AI's outcomes, and to identify and correct any issues or problems. You can also report and disclose your AI's outcomes to your stakeholders and regulators, and to learn from your experiences.

添加您的观点

加载更多内容

6 Review your practices

AI is not a one-time or isolated activity. It is part of a broader and dynamic ecosystem that involves people, processes, and systems. Therefore, you need to review your AI practices regularly, and reflect on their strengths and weaknesses. You can use methods such as self-assessment, peer review, or external audit to assess and improve your AI practices, and to align them with the best practices and standards in your field. You can also share your AI practices and learnings with other AI practitioners and communities, and to contribute to the advancement and innovation of responsible AI.

添加您的观点

7 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?

添加您的观点

Zaki E.

Data & AI Leader at Electronic Arts (EA) ?? | Designing Scale & Innovation ??
举报内容
We need to tackle some of the challenges upon releasing any AI product: - Toxicity, involving harmful behaviour or outputs to customers or communities safety. - hallucinations and reliability where AI produce false or baseless information. - intellectual property and protection. There are some mitigation strategies that include curating training data, transparency, employing guardrail models, diverse human annotators, and AI literacy for stakeholders. Building and using AI responsibly requires defining specific use cases, assessing risks, iterative development, enforcing accountability and most importantly establishing an AI Ethics framework.

已翻译

赞

Artificial Intelligence

+ 关注

给文章评分

我们借助人工智能创建了此文章。您认为这篇文章怎么样？

很棒不太好

举报此文章

查看全部

How can you deploy responsible AI?

1

2

3

4

5

6

7

1 Define your purpose

2 Assess your risks

3 Engage your stakeholders

4 Implement your standards

5 Monitor your outcomes

6 Review your practices

7 Here’s what else to consider

Artificial Intelligence

给文章评分

感谢您的反馈

更多Artificial Intelligence相关文章

更多相关阅读内容

How can you deploy responsible AI?

1

2

3

4

5

6

7

1 Define your purpose

2 Assess your risks

3 Engage your stakeholders

4 Implement your standards

5 Monitor your outcomes

6 Review your practices

7 Here’s what else to consider

Artificial Intelligence

给文章评分

感谢您的反馈

查看其他技能