Technology | Innovation

Frontline employees need to be involved in development and deployment in AI

Involve front line staff to ensure the ethical and fair development and deployment of AI and automation, Community and Public Sector Union Senior Policy Officer, Osmond Chiu has told a CEDA audience in Adelaide.

Speaking at an AI and ethics series event, Mr Chiu said we need to “empower those who know what’s going on in the workplace.”

“We need to involve front line staff in the development and deployment in automation and AI because they can often see the perverse impacts and the unforeseen outcomes of decisions.

“Workers have to be at the table, they have to have a say about how these changes are implemented and managed through consultation, negotiation (and) sharing information to adjust to the impact.

“Ultimately AI is created and shaped by humans – our values, our flaws our biases are encoded.

“We need human staff who understand the decision-making process, keeping track.

“The benefits of AI can be shared but only if transitions are managed well.

“A situation where business gets all the rewards and the risk falls entirely on workers will not be palatable to the community.

“We also need the strength and the ability of people to challenge decisions made by AI.

“There needs to be protections for conscientious objectors and ethical whistle blowers (too).

“This discussion we have about AI and ethics has to be embedded in the real world and deal with outcomes not just processes.  

“We can’t overlook questions of ownership and distributed fairness.

“There is growing research that inequality is bad for social and economic outcomes and AI and automation has the potential to worsen it even if we have an ethical approach –  for example while governments can provide better datasets to minimise biases, the use of AI will create new data.

“Transparency around AI becomes even more important when it’s used for complex assessments.

“Within my sector, the public services, there’s already increasing automation of service delivery and the push in self-service online as well as the increasing use of data matching.

“Within Australian Public Service (APS) as an example, we’re seeing the automation of tasks but also the increasing use of the AI bots – virtual assistants used by Human Services, the ATO, the NDIA.

“The APS is also establishing the Augmented Centre of Excellence to encourage the greater use of these assistants, and there’s even plans for physical robots to complement staff at Centrelink offices.”

Also at the event was AlphaBeta Principal and Melbourne Office Leader, Dr Jim Minifie, who spoke on the future of work.

“The current pace of structural change is not really higher now than it has been in the past,” he said.

“What you are seeing though…is a shift towards high skill, high touch occupation and tasks.”

Dr Minifie said these included community and personal services.

“When you ask employers now compared to a few years ago, they’re asking for digital literacy, critical thinking, creativity but also to some extent, team work, presentation skills, communication skills, more of all of those things than they did in the past,” he said.

“You can think about those capabilities as being complements to the raw computing and the communication and prediction inference powers of AI.

“Similarly, the task mix is shifting in the same direction towards interpersonal creative decision making.

“Our estimate is between 2000 and 2030, the average work week will shift about 10 per cent…towards those more personal and creative decisions.

“You can think about that as 10 per cent of the workforce changing entirely, but actually a lot of it happens at the task level, not at the job level.

“Our estimate is that about three quarters of that change happens within jobs, rather than through people shifting jobs.

“(And) those tasks pay better. One estimate…shows they pay about 20 per cent higher than the more automatable tasks.”

This enables a shift to more ‘artisan’ production he said, with machines doing routine tasks enabling people to do more human and creative activity.
This event is part of a national AI and ethics series CEDA has been hosting across Australia. To hear from international speaker Microsoft President Brad Smith on security and privacy in an AI world, watch the video below.