Where are we at with human interaction with AI? What impact in terms of ethics does this have at the moment? These were the focal topics of investigation during a panel discussion at the Digital Transformation EXPO Europe 2021.
Moderating the 'AI Ethics in Practice: How to Turn AI Principles Into Practical Governance and Compliance?' panel was Sherin Mathew, CEO and founder of Innovation Exchange, who opened the panel by discussing AI through the lens of regulation, particularly how AI regulation can become governmental policy.
Tim Clement-Jones, co-chair of the All Party Parliamentary Group of Artificial Intelligence, emphasized that accepting the need to put AI regulation into policy is "one the most important points to this question." "We are at a crossroads," he said. With the EU and the US already making inroads, the UK must ensure that it doesn't fall behind. The EU approach "appears to be the best model," he claimed.
Minesh Tanna, global AI lead of law-firm, Simmons & Simmons and chair of Society for Computers and Law (SCL) AI Group concurred with Clement-Jones, praising the EU draft on AI regulation as "very good since it is detailed." Ensuring that any policy is detailed will avoid significant "pitfalls."
Another central theme of the panel concerned what steps organizations can follow to ensure that they are regulation-ready. Sara El-Hanfy, innovation lead, machine learning and data at Innovate UK, brought attention to investing in workforce skills. "Employees must have lifelong learning around AI" she claimed. Simon Greenman, partner and member of the World Economic Forum's Global AI Council, told the audience that organizations must figure out what AI they are using. Almost all attempts to regulate AI miss the critical question regarding what the AI being used is. "Figure out what AI you are using," he stressed.
The panel discussion quickly evolved into a debate about the benefits of regulating AI. Greenman argued that the competitive advantages of AI regulation would be "immense," a view shared with Tanna, who remarked that contractual assurances are now more commonplace. In addition, ethical business practices as a paradigm was raised, with more customers choosing to shift to ethical consumerism. This point was highlighted by El-Hanfy, who highlighted AI regulation as an "immediate ethical opportunity."
The final question of the session concerned best practices for businesses to get regulation-ready. Most of the panelists were in agreement with Greenman, who delineated six practices that included:
- Getting those at the board level involved in the conversation
- Ensure effective teamwork
- Emphasise leadership visibility
- Don't forget about controls
- Use explainable language
- Be sure to have an AI "kill-switch"
Wrapping up the session, Clement-Jones highlighted digital literacy. One of the most significant pitfalls in achieving AI regulation, he said, is getting those on the board level to have a better digital understanding. "Time and time again, I see this problem," he rued. "This is no good unless those sitting on a board are AI literate. AI regulation is not an issue for the future; it is an issue for the present."