I’ve frequently made the case for why large companies need a more bold and strategic approach to ethical AI in order to match the pace of AI innovation. This is #3 in a series of posts on Enterprise Artificial Intelligence (AI) that takes a closer look at how global companies are currently integrating ethics into their AI programs, processes, products, infrastructure, and even culture.
[Here are links to blog post #1 “Welcome to the Age of the Ethical Consumer” and blog post #2 “People Should be at the Heart of AI Design Process”, in case you missed them]
I was a guest at the Atos technology days co-located with VivaTech in Paris, France’s premier show for startups and innovation so I took this opportunity to ask John Hall, Head of Strategy & Portfolio at Atos, what role ethical AI plays in the global company’s strategy and get his perspective on the best way to balance innovation (in AI) with ethics.
Trending AI Articles:
1. Deep Learning Book Notes, Chapter 1
2. Deep Learning Book Notes, Chapter 2
3. Machines Demonstrate Self-Awareness
4. Visual Music & Machine Learning Workshop for Kids
The biggest fear for many enterprises is that Ethics, in the form of Diversity and Inclusion, will stifle innovation but it’s heartening to hear from Hall that it is possible to do both. In fact, being ethical, increasing diversity, large organizations can help accelerate the adoption of AI. He also emphasizes the need for explainable, auditable AI and ethics by design.
It’s not always easy to find partners who align with your company’s values so in the absence of consistent standards for an emerging area like ethical AI, Atos has taken the lead by defining its own “Ethics by Design” policy, which will serve as guidance for their stakeholders and partners.
More from our Q&A below.
Mia Dand: What role does AI play in Strategy and Innovation at Atos?
John Hall: We consider AI as one of the main sources of digital disruption and therefore it will inevitably be a cornerstone for innovation for Atos and our customers. Properly implemented, AI based technologies can bring all manner of benefits from cost savings through automated process efficiency; to new value creation through previously unimagined services; and increased customer satisfaction through more contextualized and predictive engagements.
The definition of the scope of AI varies significantly — for some it includes everything from Predictive and Prescriptive Analytics, to Robotic Process Automation, Machine Learning and Deep Neural Networks. Atos employs flavors of all of these technologies in seeking to bring new and exciting capabilities both internally and to its customers.
But AI also changes the game in the way the value is perceived and delivered — inevitably this must influence business strategy. In some areas the need for responding to this change will probably be quicker and further reaching than we might expect. Those that fail to respond appropriately could find themselves quickly losing market relevance.
MD: How do you ensure that the use of AI by Atos is ethical and responsible?
JH: With all the new and exciting opportunities presented by AI, we have to make sure we use them properly and responsibly. Most things in life can be exploited for good or not so good purposes and can have both positive and negative knock on wider implications. The human race has already gone through a number of industrial and technological revolutions, all of them in turn bringing elements of fear of the unknown and disruption to the “conventional” way of doing things. Historically governments, businesses and societies have not always done a great job in recognizing the impact of technological advances and have had to play “catch up” as a result. The current increased focus on CSR, data privacy and sustainability is clear evidence of this.
We have an opportunity to learn from history and proactively deal with the potential challenges of the digital revolution. This is why we advocate “Corporate Digital Responsibility”, a natural progression from Corporate Social Responsibility that recognizes the new and different risks and challenges that are present in the way we use Digital Technologies.
MD: What criteria do you use to evaluate if partners are ethical/responsible in their use of AI?
JH: This is an area that still requires some work because the transparency of supply chains is not yet where it needs to be. The ideal situation is one of working with partners who share the same value set as you do, but proving this to be the case, particularly in the new and emerging area of AI is easier said than done. That is why Atos is committed to defining its own “Ethics by Design” policy that will become a blueprint for how we operate and will help set the standards of engagement that we expect our partners to work to. It will address the methods, tooling, culture and governance that will shape our enterprise DNA. It will hopefully become a set of standards that both partners and customers are able to make use of.
MD: How should companies balance innovation with AI and ethics?
JH: The starting point should be for a business to have a clear position on its values and underpinning ethical principles. These should then be the moral compass which guides the implementation of new technologies. For example, if a business agrees that Diversity and Inclusion is a core value, then related AI driven processes should avoid any negative bias.
This does not imply that innovation will be stifled as a result, quite the opposite, it means that businesses can focus on implementing innovations in ways that are far more likely to be accepted and therefore deliver against expected outcomes. Failure to respect the “boundary of acceptability” will likely lead to rejection of innovations and, consequently, wasted investments. It is important to recognize that the boundary of acceptability is not necessarily fixed, it can be influenced by building strong user engagement and trust. It is important for businesses to demonstrate that they are acting in a trustworthy and transparent way if they want to push technologies that users may otherwise be unsure about. This is why explainable and auditable AI and Ethics by Design is so important.
And it’s a wrap!
Stay tuned as I continue to cut through all the hype and jargon to bring you the inside scoop on how enterprise companies are ensuring AI is being deployed ethically in their organizations. Tweet me @MiaD or leave a comment below as I would love to hear about your AI trials and tribulations!