Ethics & Technology: Does AI Need a Pilot at the Helm?
Summary: Advances in conversational AI technology necessitates careful balancing of interests that foster innovation while mitigating potential harmful impacts.
The Allure of Conversational AI
The launch of ChatGPT-3 has captured unprecedented interest with over 100 million active users per month.
Consumers and organizations alike recognize just how advances in conversational AI can potentially transform virtually every aspect of our personal lives and professional interactions. ChatGPT-3’s advanced application of deep learning generates convincing human-like conversations with impressive results. In recent tests ChatGPT-3 passed U.S. Medical Licensing Exam and an MBA exam from one of the most prestigious business schools.
The business benefits of conversational AI are immense so much so that Microsoft has invested $ 10 billion to Open AI to accelerate commercial applications of ChatGPT-3. Microsoft intends to “supercharge” its products with the conversational capabilities of ChatGPT-3.
But lets’ take a deep breath before we let such transformative technology get ahead of ourselves. Its potential utility notwithstanding, ChatGPT-3 is still in nascent stage with, admittedly many imperfections. Our exuberance may prove to be premature. In the absence of meaningful analysis an unfettered adoption of conversational AI without regulatory guardrails and human oversight can lead us into a dystopian world, foreshadowed by the English novelist Aldous Huxley. Too much reliance on technology without understanding of its socio-economic impacts can have grave, albeit unintended consequences, such as employment, education, inclusivity and fairness.
After all, while Chat-GPT-3 is a step forward, its “cognitive” capabilities are based on pre-existing data sets, with built in biases which can be amplified resulting in a myriad of potential harms. ChatGPT-3 has a “dark side” that ought to be cause for concern and push the pause button before further proliferation of such consequential technology.
Risk or opportunity?
There needs to be a balancing of the risks associated with the application of conversational AI and its potential opportunities. If deployed in a thoughtful manner, with meaningful regulation of AI that considers a risk-based approach with robust oversight then the potential benefits of conversational AI may be realized. It should be deployed as a human centered partnership with conversational AI applications that enrich the workplace experience. Like all the innovations that have gone before it, AI has the potential to create the need for new skills and to make certain jobs more rewarding.
What is the role of the legislator?
The democratization of AI technologies implies that public authorities must take a proactive role in regulating their use. The fears associated with disruptive innovations are real. Despite its prowess, artificial intelligence remains dependent on humans, whether to use it or to regulate it. There must be indeed a pilot at the helm, we just don’t know exactly which course he is heading for yet. There is increased momentum for comprehensive AI regulation, most notably, the draft EU Artificial Intelligence Act. With the potential ubiquitous adoption of conversational AI, such as ChatGPT-3, the momentum for more rigorous regulation and governance is growing with the EU to ensure that measures are instituted that would prohibit specific application of conversational AI that is deemed to be infringing on fundamental human rights.