AI Unveiled: Challenging Misconceptions and Embracing the Future

Is using AI cheating?

Manca Vitorino

A colleague recently told me she has been using RiskGPT to improve the structure of her emails and communication. As a non-native English speaker, she found that using a chatbot to improve sentence structure and syntax made her emails clearer and more impactful. “I need to be careful though or people will notice,” she told me.

The unnecessary but very human concern made me wonder whether 130 years ago, people felt similarly about the dishwasher, invented by Josephine Cochran. Were they concerned that rather than efficient, they would appear slovenly. It took another 70-80 years, until roughly mid 1960s, before dishwashers became a household staple with uptake largely attributed to technology becoming more generally available because of other innovations such as furnaces, which can produce enough hot water to run a dishwasher, for example.

Today - as far as I know - nobody is wracked with feelings of guilt when we use the dishwasher now and we fully accept that it can do a better job than us, while also using less water. So why do we feel differently about AI?

While the pace of technological advancements has been speeding up, until 2022, AI remained in the domain of subject matter experts, largely inaccessible to the broader population. OpenAI changed that dramatically when ChatGPT was made available – for free – to the public in late 2022 and it had the potential to change everything. We clearly did not need to wait 70 years between invention and broad adoption but despite the rapid spread of chatGPT (or Bard, Copilot, Claude etc) users, many remain sceptical over the benefits of its application.

The way people adapt to change has been well documented. The change curve lists stages such as shock and denial, followed by anger and fear, acceptance and finally commitment.

Some of the guilt we experience when using new and more disruptive technologies, such as AI, might stem from the fact that we might be further along the change curve, fearing the reaction from those still in denial or fear stages. More likely though is the fact that AI can do things that until recently, only humans could.

So trusting AI with tasks we prided ourselves on being able to do, likely taps into insecurities around our value and contribution.

But what AI cannot do is create purpose. Each time we interact with AI or GPT chatbot, we provide the problem, question, and the purpose. Using a tool that can help solve the problem is efficient, innovative and creates the value, which AI cannot do on its own. None of that is cheating.

Is my data safe with AI?

Ashley Byford-Bates

AI is a product of human design and reflects the values, priorities, and potential biases of its creators. While AI has the potential to transform various business functions, it should be seen as a tool that aids in support roles rather than a direct replacement for humans (hence the “co-pilot” branding).

When integrating AI systems, protecting intellectual property and ensuring data security are paramount concerns. Measures such as restricting data exposure, implementing encryption, secure storage, and stringent access controls can help mitigate risks like data breaches and misuse.

To ensure responsible use of AI, clear guidelines on data usage, storage, and user consent should be established in the business strategy. It is also important to consider the risks associated with Large Language Models and take steps to protect customers while they engage with new capabilities.

Dealing with the increasing complexity of AI systems requires a balanced approach between innovation and a level of caution. Some of these considerations were outlined in the Data Literacy course:

- Robust Data Governance: Establish strong data policies to minimize biases and safeguard data integrity.

- Transparency and Explainability: Develop AI systems that are clear in their operations and decisions to ensure accountability.

- Regular Audits: Frequently audit AI systems to detect biases, errors, and ensure compliance with legal and ethical standards.

- Ethical Guidelines: Promote ethical development and usage of AI based on principles outlined in RELX Responsible AI Principles.

- Continued Human Oversight: Maintain active human involvement in AI operations for addressing unexpected issues.

In the commodities sector specifically, there is significant promise for utilising AI, some examples include:

 - Market Analysis: Use AI to analyse data for predicting market trends which can aid strategic trading decisions.

- Supply Chain Optimisation: Employ AI to enhance logistics management by optimizing inventory levels leading to improved efficiency.

- Sustainable Practices: Implementing AI can monitor activities like mining or agriculture to minimize environmental impact through optimized resource use.