AI and the ascent of humankind

“Just don’t try to fit a square peg into a round hole.”

Whilst the possibilities of Artificial Intelligence to create better futures are no doubt countless, there are important unanswered questions relating to the ethics and unknown risks of its adoption in business.

Curious to lead the debate on how organisations approach this topic, we gathered the brightest leaders to discuss People-first AI, where company leaders from Google, Bupa, BT, Zopa and The Economist shared learnings and challenges on the hot topic. Here are some conversation headlines:

• Iterative testing minimises risk. Understanding AI’s opportunities and untapped potential will only be realised if leaders are purposeful in their approach. Experimenting and testing in an iterative manner will derisk the implementation of new technologies and offer tangible measurable efficiencies that can be rolled out across organisations.

“We’ve taken a five-step process to three steps. Yes, it’s trial and error, but we’re testing and learning as we go, which means we’re building a case for what works for our business and why.”

• “I’m constantly being asked — give me a use case for AI”. Tools can augment processes, and we’re seeing many successful implementations of AI to increase efficiencies in mundane work. This frees up valuable time for people to focus on what people do best; injecting creativity, wisdom, and those all-important moments of collaboration.

“We’ve trained the tool in our brand tone of voice, which is useful for mass content creation. Ultimately someone needs to check and often modify that content, maintaining the brand personality.”

• Inclusive design principles will mean we avoid biases and unintended consequences in AI. There are reports of tools behaving in a racist and misogynistic manner. This is often due to unconscious bias from inaccurate data sets, greed or simply a poor development process.

“We must follow human-centric principles: Don’t design for yourself. Don’t mark your own homework. Don’t post-rationalise the solution that you always wanted it to be. Include diverse users that represent the population in design.”

• Like any disruptive technology, the unknown is shrouded with anxiety as well as excitement. The media paints a picture of a labour market that fears substitution with clickbait headlines that AI will take our jobs. Artists and makers have already taken action against ignored intellectual property rights, and teachers are grappling with issues of plagiarism and authentication. The answer? One contributor said: “Let’s create principles for engagement. These can and should be reviewed over time, but this will mean we have a framework for how we use and adopt new tech in a safe and measurable way. We need to tackle fears head-on, we can’t ignore them, and we can be transparent in our approach.”

“Reports show that 43% of workers use Chat GPT in the workplace, and 63% do so without their manager’s knowledge.”

• “AI is great looking back, but not looking forward.”. Chat GPT, made by Open AI, uses deep learning techniques and massively large data sets to understand, summarise, generate and predict new content. Large Language Models, therefore, tend to be past-focused, and only as valuable as the data it has access to; making forecasting very difficult.

“If we don’t experiment with opportunities presented by AI we will become superfluous.”

• “Risk Teams need to be accountable, now”. Guardrails and guidance on how to use AI will be essential in order to protect people and businesses, but we can’t wait for legislation to come into place. Risk Teams have to be on the front foot, setting standards and shaping expectations for use by both businesses and consumers.

“AI is already making decisions that directly impact human lives. We can’t wait to be told what’s right or wrong, it’s us that has to be accountable for its application.”

One contributor summarised the discussion; “don’t try to fit a square peg into a round hole.” Ultimately, it’s not about having an AI Strategy. Instead, we need clearly defined problems and the curiosity to explore how AI might help in solving them.

Thank you to all our contributors, and to our long-standing Exchange community. We value the space to debate and reflect with senior leaders who are leading the innovation charge in their industries, exploring new technologies and ways of working.

This discussion was hosted by our Magnetic team Jenny Burns and Jason Da Ponte. Get in touch with them at Jenny.Burns@wearemagnetic.com and Jason.daponte@wearemagnetic.com.

The last Exchange of the year will be held in our London studio in November. Follow us on Eventbrite to get the latest updates and to join the conversation.

Further reading

• If the ethics of AI is a topic that sparks your curiosity, then don’t miss the tip one of the leaders shared with us: the course “Operationalising Ethics in AI” given by Alan Turing’s Foundation.
• The 14 people who matter in UK AI policy
• The Pentagon just launched a generative AI task force
• The world needs a new Turing Test
• AI is making companies even more thirsty for your data
• AI safety summit; £13m to boost the use of AI in healthcare

Author: Laura Fernández Piñeiro