In his blog On holding back the strange AI tide, Ethan Mollick observes how the arrival of powerful AI systems like ChatGPT has been sudden and disruptive. These systems can now do many tasks previously reserved for humans, causing major impacts across work and life without warning or preparation.
Many leaders are desperately trying to eliminate the disruption of AI, hoping it won’t change anything. However, this approach is futile and poses risks. The benefits of AI will also be reduced by pretending it’s just an iteration of past technologies.
Current AI systems like GPT-4 have tremendous abilities to transform work despite being in early development. Even if progress stopped here, the impact would be substantial. Hoping to halt progress through regulation also seems unlikely.
Organizations are responding to AI in counterproductive ways:
- Ignoring it drives individuals to secretly use AI to enhance their jobs without telling leaders.
- Banning it just pushes the use of AI to personal devices while problems remain hidden.
- Centralizing it via internal systems, hoping to control and monitor use.
Centralization is the natural response but flawed for AI:
- Corporations have no AI advantage – public systems like ChatGPT are equally advanced. Private data may help but benefits are uncertain.
- Leaders rarely understand the best uses for AI across the organization – frontline workers have more insights here.
- Centralized systems feel restricted and limited, deterring innovative applications.
By treating AI like past technologies, the most transformative uses will be missed. Work must be reimagined not just automated.
To harness AI’s power, a democratized approach is needed:
- Incentives and reassurances so workers feel safe to experiment and share discoveries.
- User-driven innovation via shared prompt libraries and discussions.
- Slack time for workers to explore ideas using the best publicly available AI.
The future of work with AI remains largely unknown. Only by supporting responsible exploration can organizations hope to lead rather than follow in leveraging this extraordinarily disruptive new technology.
AIdoes.eu can help your organization with the democratized approach: a wide choice of models are available for all in your organization to experiment with safely. With the ultimate goal of simultaneously enhancing the quality of the produced work and productivity.
Summarized by Claude 2 and adapted from Ethan Mollick‘s blog (July. 23rd, 2023).
March 24th, 2024