November 07, 2024
Data-and AI- driven approach to creating positive impact:
essential, but far from being automated, ethical, safe and trustworthy
Harnessing data technology and the potential of Artificial Intelligence (AI) in a transforming world can be powerful if grounded in defined context and robust, ethical, safe and context-relevant data in order to have trustworthy and explainable outcomes
"There's definitely still room for improvement!
We're looking at how we can move forward more efficiently and intelligently. We're exploring what we can automate, where we can extract knowledge more quickly, and how we can build this into a knowledge base to create more sustainable processes. There’s still a lot of controlling human brain power involved, whether it is in our beer -, or the restaurant sector, or even down to workforce planning that takes e.g. weather data into account. There are tools, but you can see that relying solely on them doesn’t work. You still need that essential component of human experience."
Katharina Kurz, co-founder and directrice of BRLO, Interview September 2024 (translated)
BRLO, founded in 2014, is a Berlin based brewery and gastronomy company that stands for modern beer diversity and distinctive experiences.
Watch an interview bite with Katharina
Artifical Intelligence (AI) ?
A series of nested concepts developed over time
1950 -
AI - The Imitation Game - a test of machine intelligence by Alain Turing
1980 -
Machine Learning (ML) - AI systems that learn from historical data
2010 -
Deep Learning - ML models that mimic brain functions
2020 -
Generative AI - Deep Learning models (foundation models) that create original content to a user’s request in form of text, images, video, audio
What is data- and AI-driven positive impact creation?
Data-and AI-driven positive impact creation is far from being automated.
An essential component is our human experience, expertise and intuition.
It is a mindset.
Related culture needs a clear vision (= a north), robust safe and ethical data governance and management frameworks, safe and ethical AI applications, a supporting infrastructure and employee empowerment (e.g. data literacy training, critical thinking, ethical approach),
- to navigate a safe and ethical usage of data technology and artificial intelligence (AI) through an iterative approach of continuously learning and adapting via AI applications grounded in specific context and context relevant data (incremental use cases)
- to decide and act with a focus on positive impact creation and following a 'beyond zero harm strategy', i.e. addressing complex challenges of the polycrises climate change, environmental crisis, healthcare, education, social justice, human rights a.o. described by the 17 global sustainable goals defined by the UN in 2015
"AI itself isn't new,
but ChatGPT's explosion brought Generative AI into the spotlight by democratizing these capabilities, putting them into the hands of everyone, with a variety of use cases primarily geared towards increasing productivity. But can we trust it for critical decisions and processes in a business context? Not just yet. The problem: ChatGPT and similar public LLMs (Large Language Models) lack context and expertise. They can spin impressive yarns but lack the depth for real-world applications and the ability to be adequately meaningful in a specific domain or customer context. And, with widespread examples of data hallucinations as well as risks of data poisoning and fraud, they also lack the trustworthiness needed to drive real adoption by business users and leaders."
BUT
Why does data-driven positive impact creation matter?
Is it an opportunity to embrace change and how is the engagement today?
Which hidden barriers and underlying dynamics are impeding a safe, ethical and trustworthy data-driven positive impact creation?
What can we do to overcome them?
Why does it matter?
Data- and AI-driven impact creation reinforces our ability to address complex issues effectively, equitably, and sustainably. It enables a proactive approach to pressing challenges, fostering innovation, and enabling a way forward to bring trustworthy outcomes and positive impact. However, harnessing data technology and the potential of AI needs to be understood for what it is: Supporting tools. Robust, safe and ethical data governance - and management frameworks, process quality and infrastructure quality, data literacy, critical thinking, creativity, communication and cooperation for trustworthy outcomes, decisions and actions is our responsibility as humans.
Opportunity to embrace change?
Integrating trustworthy outcomes of safe and ethical AI applications that are grounded in specific context and context related data, in the business strategy, - models, value propositions and processes help to navigate a foggy journey of transformation, keep us from venturing too far down a dead end and help to embrace pivotal shifts:
- from degenerative to regenerative business models
Sustainability and beyond: Data and AI usage earmarked for driving sustainable and regernerative practices, reducing environmental footprints (net-zero), enhancing resource efficiency and increasing the environmental and societal hand print (net-positive).
- from treating symptoms to addressing causes
Data-driven positive impact creation addresses complex challenges related to the 17 UN Sustainable development goals due in 2030
- from short term profit to long term regenerative and societal value creation (=positive impact)
Innovation supported by safe and ethical AI applications contributing to long-term environmental and societal benefits, ultimately considering business as an integral part of nature and society.
- from shareholder to stakeholder approach
Equity: Ensuring that robust, safe and ethical data foundation, - AI applications and trustworthy outcomes are accessible, transparent and of benefit to all, reducing disparities and promoting inclusiveness
New regulation and standards and prioritization of ethical AI principles, transparency, and responsible AI development is needed to unlock the potential of AI for positive impact creation.
Engagement today
Climate action
Data-driven models are supporting climate predictions and helping design effective mitigation strategies
- IBM’s Green Horizon project in Europe has enabled a 30% reduction in air pollution in cities like Copenhagen by 2023" (IBM, 2023)
- Google’s DeepMind AI has improved wind farm energy output by 22% in 2024 by optimizing predictions in European renewable energy projects" (DeepMind, 2024)
Engagement today
Health care
AI-powered diagnostics and predictive analytics have revolutionized healthcare, leading to early disease detection and personalized treatments.
- Google Health’s AI system for breast cancer detection has shown significant improvements in accuracy over traditional methods, with a 10% higher accuracy rate in 2023" (Google Health, 2023)
- The NHS AI predictive analytics program has the potential to prevent over 22,000 deaths annually by identifying high-risk patients earlier" (NHS, 2023)
Engagement today
Education
Adaptive learning platforms powered by AI have improved access to education and personalized learning experiences.
- Knewton, an AI-driven education platform, has improved student learning outcomes by tailoring content to individual learning speeds, leading to a 25% increase in student performance across Euopean schools. (Knewton, 2023) .
- Coursera, which now has over 100 million learners worldwide as of 2024, uses AI to provide personalized course recommendations, with significant growth seen in European countries like Germany and the UK.
Engagement today
Smart cities
The integration of AI in urban planning led to more efficient energy use, better traffic management, and enhanced public safety.
- Barcelona's AI-driven water management systems saved €50 million in 2023, reducing water consumption by 30%" (City of Barcelona, 2023)
- Amsterdam's AI traffic management system cut congestion by 35% in 2024, reducing travel times by 20%" (City of Amsterdam, 2024)
"This is what we expect to see in 2024:
Companies increasingly wrestling with the balance between doubling down AI and digital investments and stepping up their focus on minimizing their environmental impacts."
For building a culture of data-driven positive impact creation it is crucial:
- to establish a clear vision and communicate the value of data-driven approaches to all stakeholders
- to navigate carefully and thoughtfully from the early beginning, not avoiding extra miles to build safe, ethical and accountable AI applications and avoiding digital harm
- to build customized safe and ethical AI applications trained on specialized data with checked quality of data, processes and infrastructures as well as trustworthy outcomes: iterative approach and incremental use cases
- to invest in data infrastructure and/or tools that are ideally not part of surveillance business models but independent alternatives
- to empower stakeholders of the ecosystem to analyze, interpret and apply it
- to support an open mindset, promoting data literacy and providing training and resources for employees to enhance their data and 4 C (critical thinking, creativity, communication and cooperation) skills
Data-driven positive impact creation
Informed decision making
By analyzing big data sets, organizations from the public, private and non-profit sectors can make more accurate and smarter decisions if the data and approach is earmarked. Resulting in more effective interventions, resource optimization and better policies for complex challenges addressed in the 17 UN Sustainable goals. Pre-requisites: Building and using context-related, ethical and safe data foundation and AI applications. To avoid overrelying on a data-driven approach: 4 Cs, human expertise, experience and intuition need to be fostered.
Sustainability and beyond
Data centers currently consume estimated. 200 TWh electricity per anno. 3 to 5 Million gallons of water and contribute to 4% of green house gas, more than the aviation industry. The capacity is expected to triple till 2030, the year the UN SDG are due. On the other hand data and AI driven innovation can be part of the solution to manage resources better (e.g. optimized hardware consuming less energy, energy efficiency coding) and protect the environment by predicting and mitigating issues like e.g. climate change.
Pre-requisites: Building and using context-related, ethical and safe data foundation and AI applications. To avoid overrelying on a data-driven approach: 4 Cs, human expertise, experience and intuition need to be fostered.
Accountability and transparency
Harnessing data and AI applications need to be grounded in specific context and robust, ethical and safe data foundation to ensure trustworthy and explainable outcomes that increase transparency and enable to hold organizations in the private, public and non-profit sectors accountable for their actions and related impact.
Pre-requisites: Building and using context-related, ethical and safe data foundation and AI applications. To avoid overrelying on a data-driven approach: 4 Cs, human expertise, experience and intuition need to be fostered.
Innovation
Data and AI-driven innovation are opening new avenues for creating positive impact and supporting pivotal shifts, job creation, and technological advancements.
Pre-requisites: Building and using context-related, ethical and safe data foundation and AI applications. To avoid overrelying on a data-driven approach: 4 Cs, human expertise, experience and intuition need to be fostered.
Diversity I Equity I Inclusion
Earmarked applied , data-driven initiatives can identify and address disparities across different populations, leading to more equitable distribution of resources and opportunities. This is crucial e.g. in education, public health,climate action where data can reveal gaps that might otherwise be overlooked.
Pre-requisites: Building and using context-related, ethical and safe data foundation and AI applications. To avoid overrelying on a data-driven approach: 4 Cs, human expertise, experience and intuition need to be fostered.
Scalability and efficiency
Data- and AI- driven approaches support thoughtful decision-making on scalable solutions which can be deployed broadly and quickly. Earmarked AI algorithms, if trained on robust and context related data foundation, can process big datasets at speeds impossible for humans, increasing the efficiency of rapid responses to e.g. emergency crises like natural disasters.
Pre-requisites: Building and using context-related, ethical and safe data foundation and AI applications. To avoid overrelying on a data-driven approach: 4 Cs, human expertise, experience and intuition need to be fostered.
Predicitive potential
Predictive analytics can anticipate future trends and outcomes, enabling proactive rather than reactive approaches. E.g. in healthcare, predictive models can identify at-risk populations before diseases become widespread, potentially saving lives and reducing costs.
Pre-requisites: Building and using context-related, ethical and safe data foundation and AI applications. To avoid overrelying on a data-driven approach: 4 Cs, human expertise, experience and intuition need to be fostered.
Hidden barriers and underlying dynamics impeding data-driven positive impact creation
"I see AI as born out of the surveillance business model . . . AI is basically a way of deriving more power, more revenue, more market reach,” she says. “A world that has bigger and better AI, that needs more and more data . . . and more centralised infrastructure...The [AI] market is crazy right now. Seventy per cent of Series A [early-stage start-up] investment is coming from the hyperscalers, and the majority of that goes back to the hyperscalers,” she says, referring to cloud companies Microsoft, Google and Amazon. “It’s like a Potemkin market, it’s not a real market..."
Data privacy and cybersecurity
- In 2023, ransom ware affected 66% of all organizations as reported by Sophos
- A 2023 EU survey reveals '60% of Europeans think getting control of one's own data and digital legacy is not well protected in their country'. AI index reports that 'people across the globe are more cognizant of AI’s potential impact—and more nervous'. Trust is low.
- The GDPR (General Data Protection Regulation) was published in 2016 by the EU but organizations are still facing challenges with compliance.
Ethical issue in AI development: Bias, transparency & explainability, privacy and autonomy
- Bias: AI can perpetuate human biases, leading to questionable outcomes/ recommendations. Bias in data collection and processing worsens the situation
- Transparency & explainability: Trust in robust, safe, ethical input data, ethical AI applications and trustworthy and explainable outcome is key but in reality we a far from being transparent and trustworthy
- Privacy: The main concern is personal data handling (see data privacy above) posing risks like intentional breaches or accidental leaks that might lead to identity theft, fraud and other types of abuse
- Autonomy: AI applications are able to either hinder or support human autonomy. And we have ethical challenges posed by autonomous systems (e.g. vehicles and weapons)
E.g.
A 2020 Harvard study highlighted that facial recognition algorithms boast high classification accuracy (over 90%), but these outcomes are not universal. A growing body of research exposes divergent error rates across demographic groups, with the poorest accuracy consistently found in subjects who are female, black and 18-30 years old.
Due to mistaken identity and privacy invasion facial recognition is already regulated in use (EU AI act) or banned e.g. Belgium and Luxembourg.
Unchecked data foundation and AI applications
- Poor data governance and management frameworks, poor data processing (collecting, manipulating and interpretating) and/or poor infrastructure in terms of storage, security and supervision undermine a robust, context -related, ethical and safe data foundation that is the opposite of an unchecked data avalanches.
- Data and AI applications built as part of value propositions deriving from a surveillance business model lack privacy, accountability and trustworthiness
- For explainable and trustworthy outcomes earmarked AI applications need to be grounded in context and trained w context-related specific data.
E.g.
AI index reports that robust and standardized evaluations for LLM responsibility are seriously lacking
Sustainability and beyond
- Data centers currently consume est. 200 TWh electricity per anno, 3 to 5 Million gallons of water and contribute to 4% of green house gas, more than the aviation industry. The capacity is expected to triple till 2030, the year the UN SDG are due.
- On the other hand, data and AI driven innovation can be part of the solution to manage resources better (e.g. optimized hardware consuming less energy, energy efficiency coding) and protect the environment by predicting and mitigating issues like e.g. climate change.
AI and the Global South
- In 2023, it was reported that only 12% of global AI investments reached developing countries, indicating a widening digital divide that hampers equitable AI deployment. (World bank, 2023)
- While many of today’s 'top' artificial intelligence (AI) models are designed in the US (2023: 61), EU (2023: 21) and China (2023:15), much of the labour that fuels the hype is based in the Global South, raising questions over who stands to gain from the technology and at what cost. From labour conditions to democracy and the environment: Why countries need to move beyond “catching up with the North” when deciding what role AI can and should play in their societies:
Being overly reliant on data- and AI-driven approaches is risky
AI index reveals that "AI beats humans on some tasks, but not on all. AI has surpassed human performance on several benchmarks, including some in image classification, visual reasoning, and English understanding. Yet it trails behind on more complex tasks like competition-level mathematics, visual commonsense reasoning and planning".
Regularity challenges
As of 2024, 75% of European countries are still developing comprehensive AI regulations, with the EU’s AI Act expected to be implemented by 2025 to address these gaps.
The number of AI-related regulations in the U.S. has risen significantly in the past year and over the last five years. In 2023, AI index reports of 25 AI-related regulations, up from just one in 2016. Last year alone, the total number of AI-related regulations grew by 56.3%.
AI development - today and in the future
Today, only a few leading tech firms/developers -mostly from the Global North- shape AI development. Their AI systems are ultimately used for making impactful determinations in all sectors including the sensitive ones like health, energy, jurisdiction, military and financial services. This almost monopolistic status with its concentrated economic forces goes against the real need in complex context: Healthy ecosystems aiming to create positive impact in a collaborative approach with trusted partners and stakeholders
"We Need To Rewild The Internet
The internet has become an extractive and fragile monoculture. But we can revitalize it using lessons learned by ecologists"
Read the essay by Maria Farell and Robin Berjon here. 16.04.2024. Neoma.
Hidden barriers and underlying dynamics are powerful so what can we do to overcome them?
"This sort of transformation does not happen overnight. Businesses must have a strategy to adopt data-driven decision-making practices. Focusing on specific metrics and key performance indicators helps measure success and keeps companies from venturing too far down dead-end roads. Far from being automatic, it requires a clear vision, supporting infrastructure and employee empowerment to be successful. It also requires an eye to the risks, which must be mitigated from the start."
Checked robust, safe and ethical data foundation and infrastructure
Data governance and management frameworks, data processing (collecting, manipulating and interpretating) and/or infrastructure in terms of storage, security and supervision are key elements to create robust data foundation for data and AI-driven innovation.
E.g.
The General Data Protection Regulation GDPR remains a global model for data protection, and in 2023, the EU introduced additional guidelines to ensure AI applications align with privacy standards, further strengthening responsible data use in Europe.
Accountable AI development
- With the integration of data- and AI-driven approaches in business activities, data privacy and cyber resilience must go beyond preventing data breaches and maintaining business continuity: Advanced threat detection, machine learning, adaptive behavioral analytics, organizational policy and process upgrades are moves forward.
- AI is mainly designed in the Global North, but much of the labour is based in the Global South. Time to move beyond “catching up with the North” when deciding what role AI can and should play in their societies.
- Frameworks and guidelines are being established to ensure AI systems are transparent, accountable, and fair.
E.g.
The EU’s AI Act, anticipated to be fully enacted by 2025, is set to create a global benchmark for AI transparency, fairness, and accountability, promoting ethical AI use worldwide.
Collaborative platforms
Today, only a few leading tech firms/developers -mostly from the Global North- shape AI development. This almost monopolistic status with its concentrated economic forces goes against the real need in complex context:
Building healthy ecosystems: Partnerships between trusted stakeholders from the public, private sector, non profit and civic society are crucial for information-sharing across all sectors, shared resource development, cybersecurity preparedness planning and to promote responsible AI use. Trust will be essential.
E.g.
The 2024 AI for Good Global Summit in Geneva emphasized collaboration, bringing together 1,200 experts to develop AI solutions for sustainable development, particularly focusing on European and African partnerships.
Capacity building
Initiatives on all levels focusing on upskilling by enhancing digital literacy, critical thinking, communication , collaboration, creativity.
E.g.
The World Bank's "AI for Development" initiative expanded in 2023 to focus on enhancing AI infrastructure in Eastern Europe, providing $500 million in funding to support digital education and AI literacy. (World bank, 2023)