Industry 6.0: Navigating the ethical frontier of AI and data

The consensus from thought leaders when considering Industry 6.0 is that every operation will be controlled by human minds and performed by automated robots, combining aspects of human oversight, artificial intelligence, cloud computing, human–robots crunching big data, and quantum computing.

You might class such technological developments as revolutionary. The concept of AI, as we understand it today, began to take shape in the mid-20th century, when the foundations began with the development of logic and mathematical theories.

One such theorist was Alan Turing, who, in the 1940s, developed the concept of a universal machine that could perform computations similar to those of a computer. Turing’s proposal for a way to measure a machine’s ability to exhibit intelligent behaviour equivalent to a human is something technology companies are trying to emulate today.

While Turing could never publish his earlier Enigma work, his ideas about building computational machines were only articulated in unpublished reports at the National Physical Laboratory and Manchester University, where he worked after the war. His most famous publication, and the only one directly related to AI, is his 1950 paper for the philosophical journal Mind titled “Computing Machinery and Intelligence”.

In this paper, Turing proposed that the question, “Can machines think?” was ill-defined. Instead, he suggested an alternative: that a machine could be said to be “intelligent”. The famous Turing Test, has been a lasting contribution in the cultural debate about what it means to be intelligent.

The future impact of AI

Intelligent machines are still a futuristic idea today, yet Chat GPT4 has revolutionised our ways of working thus far. The impact that future releases of GPT5, 6, or 7 will have on individuals and developers will be revolutionary in our daily lives.

While I am excited about any new impact AI technology will have, I wonder if Alan Turing was still around today, would he be hesitant, and would the subject of ethics arise?

It has been estimated that technology will achieve pure autonomy by 2050, and to accurately predict how it will transform our lives is still uncertain. I believe it will be an interdisciplinary world coordinated by all players, from those who are part of the global political system, national governments, business (especially big tech), academia and civil society.

Yet the thought of pure autonomy by 2050 makes me uneasy, and I’m not talking about autonomous vehicles that can park my car better than me. My uneasiness lies in the bias that could be embedded into any Large Language Model (LLM).

We all have inherent biases that are a product of our environment, upbringing, and social circles. We are human, after all, but making sure these biases are not reflected in any LLM that we develop is a key step in making sure that decisions we fully automate and outsource to AI will be fair and equitable.

I’m sure in the future I’ll reap the rewards, with more timely deliveries by drone and more accurate medical diagnoses of diseases picked up earlier than ever before, but the recent announcement of the UK Courts that they have begun to consider the use of AI as an aid in Judicial decision making is a concern.

Addressing ethical concerns

Access to justice issues and benefits of digitalising the court system aside, present-day LLMs are still prone to hallucinations and bias. Having the correct level of human oversight to ensure that any technology project and the Judicial decision-making process is fair and equitable to prevent any miscarriages of justice will be essential. This is not an area of automation that society can afford to get wrong.

Industry 6.0, while theoretical, will revolutionise how we interact across all aspects of society. Technology ethics is broad and multifactorial, and I’d like to leave you with my top considerations when it comes to responsibly using data and AI within your organisation:

• Preventing bias and discrimination: data is fundamental to AI development, yet if your focus is only on data quality without considering the ethical implications, AI systems may perpetuate existing biases and discrimination.
• Safeguarding privacy over exploitation: Prioritizing ethics over a data-centric approach is essential to protect individual privacy rights. The data collection and analysis capabilities of AI are immense, but without ethical guidelines, they can lead to invasive surveillance and privacy breaches.
• Human-centric job perspectives: While AI can process data more efficiently than humans, emphasising ethics over data acknowledges the societal impact of potential job displacement. Ethical AI development involves creating a balance between automation and human employment, ensuring a socio-economically sustainable future.
• Accountability beyond data: AI systems may be data-driven, but without ethical frameworks, determining accountability for decisions made by AI becomes complex. Ethical guidelines ensure transparency and accountability in all AI decision-making, which is crucial for public trust and acceptance.
• Security and harm prevention: Data-centric approaches can overlook the broader security implications of AI. Ethical considerations encompass the potential misuse of AI, such as its use in autonomous weapons or cyber-attacks. We need to ensure AI is developed and used responsibly.
• Moral implications in sensitive industries: AI’s application in industries like healthcare, financial or legal services isn’t just about data accuracy but also moral and equitable decision-making. Ethical frameworks put in place to complement any technology project will help ensure that any automation or decision-making respects human rights, lacks biases and ensures dignity for those impacted most by any automated technological decision that is made.
• AI autonomy: While AI can operate on data autonomously, deciding the extent of this autonomy, especially in life-critical systems like healthcare or transportation, ensures safety and reliability.
• Societal impact assessment: Beyond data-driven results, it’s important to consider how AI reshapes social interactions and societal structures. Ethical guidelines from the United Nations or the EU can help assess and mitigate any potential negative impact on society.
• Integrity in the information age: The ability of AI to manipulate data and create falsified content (like deepfakes) poses a threat to information integrity. Ethical guidelines are crucial in combating misinformation and developing technical tools to spot this content. The spread of misinformation isn’t only a technology issue; increased media literacy and the consequences of misinformation spreading all must be considered in parallel.
• Environmental ethics over data consumption: The environmental impact of AI’s data consumption and processing power is a growing concern. Ethical considerations in AI development include sustainability and environmental responsibility, not just data efficiencies.

In summary, the ethical implications of AI are significant. By ensuring its development with an equitable framework at the core will ensure fairness, accountability, transparency, and respect for human rights, which are essential for the responsible use of AI in society.