Australian Government: Ethical AI Will Unlock $20T Global Opportunity

Innovation Australian Government: Ethical AI Will Unlock $20T Global Opportunity Susan Galer Brand Contributor SAP BRANDVOICE Storytelling and expertise from marketers | Paid Program Sep 22, 2022, 08:15am EDT | Share to Facebook Share to Twitter Share to Linkedin Artificial intelligence (AI) could contribute more than 20 trillion dollars to the global economy by 2030 according to the Australian government’s action plan that lays out a vision for the technology’s trusted, secure, and responsible development and adoption. To find out more, I caught up with Ian Ryan, who leads the SAP Institute for Digital Government (SIDG). He shared his perspectives on the building blocks of ethical AI based on research findings from the SIDG and the University of Queensland.

Ethical AI relies on the algorithms to collect and analyze patterns from huge amounts of information, while people have control over the final say. getty “AI represents a tremendous opportunity that cannot be ignored by the Australian government, or for that matter, organizations in the private sector,” said Ryan. “Responsible automation functions with unbiased fairness and complete transparency and accountability.

When citizens trust the government is using their data responsibly, then government has the license to innovate. ” Don’t stint on AI project resources Although software powers AI, project design, roll-out, and management is emphatically not an IT-driven endeavor. The change management aspects include revamping the culture and bringing together the right people along with significant technology resources.

In addition to data scientists, government agencies need leaders within each administrative function including legal, finance, procurement, and of course, citizen services, who understand the purpose and impact of AI-based projects. “Our research found that building representative and unbiased AI training datasets was a significant challenge for many organizations,” said Ryan. ” With close collaboration — during design and after go-live — between data scientists and the people within governmental departments who will actually be using the tools, project teams can obtain the correct data to train the AI model so it surfaces patterns and recommendations aligned with program objectives.

” Divide tasks appropriately between humans and AI models Contrary to dystopian science fiction scenarios, AI doesn’t replace the unique strengths of humans; it augments people’s abilities so they can make even more valuable contributions. Ethical AI relies on the algorithms to collect and analyze patterns from huge amounts of information, while people have control over the final say. For example, an AI-based tool can work it’s way quickly through troves of data, whether it’s ferreting out possible tax fraud or flagging early alerts to rising cases of sepsis in a community.

The technology can save the government time and costs using AI for the data collection and analysis, and human judgment to interpret the recommendations and make the best decisions. “You might have a list of 30 tasks someone performs, and decide that AI could speed up five of them, increasing efficiencies, while freeing up the balance of that employee’s time for higher-value responsibilities,” said Ryan. “People are always at the forefront and in the loop as AI collects data and sends alerts and recommendations.

An AI model can quickly spot anomalies like decreasing tax revenues, and it’s up to the human to interpret them. ” AI explainability builds trust AI explainability was among the major obstacles public sector organizations face, according to the research findings. Given the immense power imbalance between citizens and government, AI explainability would go a long way in improving public trust in government, leading to more innovation.

“When a government agency targets an individual, AI explainability provides part of the rationale for those recommendations,” he said. “Transparency in how the AI models work drives accountability, which builds greater citizen trust that the government is using their data safely and securely. People will be more willing to participate in programs, sharing their data for additional services.

” Create ethical AI guard rails Drawing from his recent research, Ryan discussed the emerging concept of having a guardrail manager assigned to assess and continuously monitor the potential impact of AI during initial project planning and design through actual results post-execution. Whether it’s one designated expert or a team of people, the guardrail manager would constantly ask questions such as what problems AI will solve, where the data will come from, how it will be used in compliance with privacy and security regulations, and what potential issues could or do arise over time. “Guardrail managers can oversee the AI tool and program to make sure it’s generating intended results,” he said.

“Drifting is possible, but with regular monitoring to review and discuss output, the organization can stay on top of potential bias issues or take advantage of unexpected opportunities AI might expose. Scaling this level of governance to prevent automation complacency will be critical as AI projects multiply across agencies. ” Ethical AI has the potential to help with “wicked problems” that people may view as too complex and difficult to address, such as climate change, domestic violence, and income inequality.

However, amazing advances are happening in AI and the prize for governments and companies that commit and act on ethical AI is tremendous. Learn more about solving complex societal problems with ethical AI. Follow me on Twitter or LinkedIn .

Susan Galer Editorial Standards Print Reprints & Permissions.